{"description":"MCP servers listed on CookiesForAI, with the facts we checked, whether each has a free tier, a link to the publisher's current pricing page (we never state prices), and the date we checked them.","license":"Our text and data: CC BY 4.0, credit cookiesforai.app. Each listed item keeps its own license.","generated":"2026-10-03","servers":[{"url":"https://cookiesforai.app/mcp/apify","slug":"apify","name":"Apify","title":"Apify MCP server: ready-made web scrapers for AI agents","summary":"The Apify MCP server gives AI agents thousands of ready-made scrapers and automation tools from Apify Store, such as Google Maps, Instagram and search result scrapers, which the agent can find and run on demand.","category":"web-search-and-scraping","publisher":"Apify","publisherUrl":"https://apify.com","official":true,"repo":"apify/apify-mcp-server","stars":9283,"repoLicense":"MIT","lastPush":"2026-10-01","registryName":"com.apify/apify-mcp-server","remoteUrl":"https://mcp.apify.com","freeTier":"free-plan","cost":"Free and open source under MIT; scraper runs are billed to your Apify account, which has a free monthly allowance, and each Store Actor sets its own price.","pricingUrl":"https://apify.com/pricing","clients":["Claude Code","Claude Desktop","Claude.ai","ChatGPT","Cursor","VS Code","Codex","OpenCode","Kiro","Google Antigravity"],"tools":["search-actors","fetch-actor-details","call-actor","get-actor-run","get-dataset-items","get-key-value-store-record","abort-actor-run","search-apify-docs","fetch-apify-docs","apify--rag-web-browser","apify--web-fetch"],"whatItDoes":["The Apify MCP server connects an AI agent to Apify, a platform of cloud scrapers and automations that Apify calls Actors. Instead of a fixed list of scraping functions, the agent can search Apify Store, read an Actor's input schema and price, and then run it. That gives it access to purpose-built scrapers for Google Search results, Google Maps contacts, Instagram, Facebook posts, e-commerce sites and many more.","Apify's server starts with two Actors loaded as tools: a RAG web browser that searches the web and returns the content of the top pages, and a web fetch tool that returns any URL as Markdown or text with JavaScript rendering. Helper tools read run results from datasets and key-value stores, check or stop runs, and search Apify's documentation. Optional tools manage saved tasks, schedules and Actor builds for people who run Apify in production.","Apify hosts the server at mcp.apify.com with OAuth sign-in or an API token, and the same package runs locally over stdio. It also accepts agent payments without an account, through prepaid tokens or x402 payments in USDC. The GitHub repo had 9,213 stars on 2026-10-01."],"useItWhen":["You need data from a site that blocks simple fetching, such as Google Maps or social networks.","You want leads, reviews or product listings as structured rows, not raw HTML.","Your agent should pick the right scraper for a new site by itself.","You already run Apify Actors and want to start, watch and read them from your editor."],"notFor":["Predictable cost: each Actor has its own price per run or per result, so check fetch-actor-details before letting an agent run scrapers in a loop.","Scraping that breaks a site's terms of use; the Actors are third-party tools and you are responsible for how they are used."],"requirements":["An Apify account (OAuth sign-in) or Apify API token","Node.js, only when running the server locally"],"connect":[{"label":"Cursor or any client with a JSON config, signing in with OAuth","command":"{\"mcpServers\": {\"apify\": {\"url\": \"https://mcp.apify.com\"}}}"},{"label":"Apify CLI: install into claude-code, cursor, vscode, codex, kiro or antigravity","command":"apify mcp install cursor"},{"label":"Local stdio server","command":"{\"mcpServers\": {\"actors-mcp-server\": {\"command\": \"npx\", \"args\": [\"-y\", \"@apify/actors-mcp-server\"], \"env\": {\"APIFY_TOKEN\": \"YOUR_APIFY_TOKEN\"}}}}"}],"connectNote":"In Claude Desktop or Claude.ai, add https://mcp.apify.com as a custom connector or pick Apify from the connector directory. To use a token instead of OAuth, send Authorization: Bearer <APIFY_TOKEN>. The configurator at mcp.apify.com builds a config with only the Actors you choose.","faq":[{"q":"Is the Apify MCP server free?","a":"The server is free, and the Apify Free plan includes a monthly usage allowance. Scraper runs past that need a paid plan, and many Store Actors charge per result (pricing page, 2026-10-01)."},{"q":"Does the Apify MCP server work with ChatGPT?","a":"Yes. Apify lists ChatGPT among its tested clients, with its own integration guide, next to Claude, Cursor, VS Code, OpenCode and Kiro."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/atlassian","slug":"atlassian","name":"Atlassian","title":"Atlassian MCP server: Jira, Confluence and Bitbucket for AI agents","summary":"The Atlassian MCP server gives AI agents access to Jira, Confluence, Jira Service Management, Bitbucket, Compass and Loom on Atlassian Cloud: search, summarize, and create or update issues and pages.","category":"productivity","publisher":"Atlassian","publisherUrl":"https://www.atlassian.com","official":true,"repo":"atlassian/atlassian-mcp-server","stars":1079,"repoLicense":"Apache-2.0","lastPush":"2026-09-15","registryName":"com.atlassian/atlassian-mcp-server","remoteUrl":"https://mcp.atlassian.com/v2/mcp","freeTier":"free-plan","cost":"Free to connect on every Atlassian Cloud plan, including Free; calls that pull enriched Teamwork Graph context use Rovo credits.","pricingUrl":"https://www.atlassian.com/platform/rovo-mcp","clients":["ChatGPT","Codex CLI","Codex Desktop","Claude.ai","Claude Desktop","Claude Code","Cursor","VS Code with GitHub Copilot","GitHub Copilot CLI","Gemini CLI","Windsurf","Docker MCP Toolkit","Kiro","Amazon Quick Suite"],"whatItDoes":["The Atlassian MCP server, also called the Atlassian Rovo MCP server, is Atlassian's own hosted bridge between AI tools and an Atlassian Cloud site. Through it an agent can search and summarize Jira work items, Confluence pages, Jira Service Management requests and Bitbucket Cloud repos, create and update issues and pages from plain instructions, and read Loom recordings, Goals, Projects and Teams data.","The Atlassian server groups its tools into permission groups per product, such as read_jira, write_jira and search_confluence, and an organization admin can grant or revoke each group. Delete and manage tools for Jira are off until an admin turns them on. To save context, the server shows agents a small set of primary tools and lets them discover the rest; adding ?tools=all to the URL returns the full list for gateways.","Atlassian runs the server only in its cloud, at mcp.atlassian.com, with OAuth 2.1 sign-in or an API token for headless use; there is no self-hosted version, and Data Center sites are not covered. Every call respects the user's existing Jira and Confluence permissions and the site's IP allowlist. The repo, which holds plugin packages and skills rather than server code, had 1,078 GitHub stars on 2026-10-01."],"useItWhen":["You want the agent to turn meeting notes or a spec into Jira work items.","You need a summary of a Confluence space or an epic's open issues without opening Jira.","You want the coding agent to read the ticket it is working on and update its status.","You want to search across Jira, Confluence and Bitbucket in one question."],"notFor":["Jira or Confluence Data Center and Server: the MCP server only connects to Atlassian Cloud sites.","Unattended write access: with write groups enabled the agent can create and edit issues and pages as you, so keep delete tools off and review changes.","Regulated workloads: Atlassian states the server does not currently support FedRAMP or HIPAA requirements."],"requirements":["An Atlassian Cloud site with Jira, Confluence, Jira Service Management, Bitbucket or Compass","A browser for OAuth 2.1 sign-in, or an Atlassian API token if your admin allows token auth","Node.js 18 or newer only for local clients that need the mcp-remote proxy"],"connect":[{"label":"Claude Code (then run /mcp in a session to sign in)","command":"claude mcp add --transport http atlassian https://mcp.atlassian.com/v2/mcp"},{"label":"Codex","command":"codex mcp add atlassian --url https://mcp.atlassian.com/v2/mcp"},{"label":"Any other MCP client: server URL","command":"https://mcp.atlassian.com/v2/mcp"}],"connectNote":"Claude Desktop, Codex Desktop, Cursor, VS Code and ChatGPT install it from their plugin or connector directories by searching for Atlassian. In Windsurf, add a custom server with serverUrl set to the v2 URL. Jira Service Management tools only work with API token auth, and only after an admin enables it. The old SSE endpoint stopped being supported after 2026-06-30.","faq":[{"q":"Is the Atlassian MCP server free?","a":"Connecting is free on every Atlassian Cloud plan, including Free, with hourly call limits that rise on paid plans. Calls that pull enriched Teamwork Graph context use Rovo credits (Atlassian pages read 2026-10-01)."},{"q":"Does the Atlassian MCP server work with Jira Data Center?","a":"No. It is a cloud-hosted service for Atlassian Cloud sites only. Data Center users need a community server, such as the MrRefactoring atlassian-dc-mcp packages listed in the MCP Registry."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/aws","slug":"aws","name":"AWS","title":"AWS MCP servers: AWS APIs, docs and infrastructure for AI agents","summary":"The AWS MCP servers give AI agents access to AWS: a managed AWS MCP Server for API calls and docs, plus more than 50 open-source servers for services like CDK, EKS, Lambda and DynamoDB.","category":"cloud-and-devops","publisher":"Amazon Web Services","publisherUrl":"https://aws.amazon.com","official":true,"repo":"awslabs/mcp","stars":9746,"repoLicense":"Apache-2.0","lastPush":"2026-10-01","registryName":"io.github.aws/aws-mcp","remoteUrl":"https://aws-mcp.us-east-1.api.aws/mcp","freeTier":"free-plan","cost":"No charge for the AWS MCP servers, and the open-source ones are Apache-2.0; you pay for the AWS resources your agent creates or uses, and AWS has a free tier for some services.","pricingUrl":"https://aws.amazon.com/blogs/aws/the-aws-mcp-server-is-now-generally-available/","clients":["Claude Code","Kiro","Kiro CLI","Cursor","Cline","Windsurf","VS Code","fx"],"tools":["call_aws","search_documentation","read_documentation","run_script"],"whatItDoes":["AWS publishes MCP servers as a family, not a single server. The awslabs/mcp repo holds more than 50 open-source servers, each for one area: AWS documentation, infrastructure as code with CloudFormation and CDK, EKS and ECS containers, serverless with SAM, DynamoDB, Aurora and other databases, IAM, pricing, support cases and more. Each runs locally over stdio with uvx and your AWS credentials, and you add only the ones you need.","The AWS MCP Server is the managed option, generally available since 2026-05-06 in US East (N. Virginia) and Europe (Frankfurt). AWS hosts it, and its call_aws tool can run any of more than 15,000 AWS API operations under your IAM permissions, with CloudTrail logging every call. It also searches and reads current AWS docs and runs Python scripts in a sandbox. Clients reach it through mcp-proxy-for-aws, an open-source proxy that signs requests with your IAM credentials.","AWS now calls the Agent Toolkit for AWS, launched 2026-05-06, the successor to the awslabs servers, plugins and skills; the repo still works and accepts contributions, and its most used projects will move into the toolkit over time. The awslabs/mcp repo had 9,746 GitHub stars on 2026-10-01. The managed server is listed in the MCP Registry as io.github.aws/aws-mcp."],"useItWhen":["You want the agent to check current AWS docs before writing CDK or CloudFormation.","You want to ask about resources in your account, such as Lambda functions or S3 buckets, in plain words.","You want the agent to debug an EKS or ECS deployment with real cluster data.","You want a cost estimate for an architecture before you build it."],"notFor":["Broad credentials: call_aws can change and delete resources with whatever your IAM role allows, so give the agent a scoped role and keep CloudTrail on.","Picking one install: the family has dozens of servers and some are deprecated, such as the Cloud Control API server, so check each server's README first."],"requirements":["An AWS account and credentials configured locally (AWS CLI profile or IAM role)","uv and Python 3.10 or newer for the open-source servers and the proxy"],"connect":[{"label":"Claude Code, managed AWS MCP Server through the AWS proxy","command":"claude mcp add-json aws-mcp --scope user '{\"command\":\"uvx\",\"args\":[\"mcp-proxy-for-aws==1.6.0\",\"https://aws-mcp.us-east-1.api.aws/mcp\",\"--metadata\",\"AWS_REGION=us-west-2\"]}'"},{"label":"Claude Code, open-source AWS Documentation server","command":"claude mcp add aws-docs uvx awslabs.aws-documentation-mcp-server@latest"},{"label":"Claude Code, open-source infrastructure as code server","command":"claude mcp add aws-iac uvx awslabs.aws-iac-mcp-server@latest"},{"label":"Any client with a JSON config, one open-source server","command":"{\"mcpServers\": {\"awslabs.aws-documentation-mcp-server\": {\"command\": \"uvx\", \"args\": [\"awslabs.aws-documentation-mcp-server@latest\"], \"env\": {\"FASTMCP_LOG_LEVEL\": \"ERROR\", \"AWS_DOCUMENTATION_PARTITION\": \"aws\"}}}}"}],"connectNote":"Configure AWS credentials before adding a server: the open-source servers read your local AWS profile, and the proxy signs managed-server requests with the same credentials. Each awslabs server README has one-click installs for Kiro, Cursor and VS Code. Set AWS_REGION in the metadata to the Region your resources are in.","faq":[{"q":"Is the AWS MCP Server free?","a":"Yes. AWS charges nothing extra for the managed AWS MCP Server or the open-source servers; you pay only for AWS resources the agent creates or uses and any data transfer (AWS announcement read 2026-10-01)."},{"q":"Which AWS MCP server should I start with?","a":"AWS recommends the managed AWS MCP Server for broad, audited access to AWS APIs and docs. Add an awslabs server when you need a specific service, such as the EKS server for Kubernetes clusters."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/blender","slug":"blender","name":"Blender","title":"Blender MCP server: 3D modeling and scene control for AI agents","summary":"The Blender MCP server, MCP for Blender, gives AI agents control of a running Blender scene: create, change and delete objects, apply materials, inspect the scene and run Python in Blender.","category":"game-and-3d","publisher":"ahujasid","publisherUrl":"https://github.com/ahujasid","official":false,"repo":"ahujasid/mcp-for-blender","stars":29830,"repoLicense":"MIT","lastPush":"2026-09-30","freeTier":"free-plan","cost":"Free and open source under MIT, and every feature works with your own API keys; optional paid plans cover the cost of AI 3D generation instead.","pricingUrl":"https://www.mcp-for-blender.com/premium","clients":["Claude Desktop","Claude Code","Codex","Cursor","VS Code","Devin Desktop (Windsurf)","OpenCode","Antigravity"],"whatItDoes":["MCP for Blender, formerly blender-mcp, is a community MCP server that connects AI agents to Blender for prompt-driven 3D modeling. It has two parts: a Blender addon that opens a socket server inside Blender, and a Python MCP server that your AI client runs. Through them an agent can create, modify and delete objects, set materials and colors, read detailed scene information, and run arbitrary Python inside Blender.","MCP for Blender can also pull in assets: Poly Haven models, textures and HDRIs, Sketchfab models, Poly Pizza low-poly models, and AI-generated 3D models from Hyper3D Rodin, Hunyuan3D and Tripo. Generation needs your own API keys on the Free plan, or a paid plan that covers the cost. The author is not affiliated with Blender.","MCP for Blender has a one-line installer that sets up uv, finds the MCP clients on your machine and installs the addon. By default it sends a minimal anonymous usage record per tool call; prompts, code and scene data are only collected if you opt in, and DISABLE_TELEMETRY=true turns it off. The repo had 29,805 GitHub stars on 2026-10-01, the most of any MCP server in our game and 3D category."],"useItWhen":["You want to block out a scene from a text description.","You want the agent to apply materials or lighting across many objects.","You need a quick prop from Poly Haven or Sketchfab placed in the scene.","You want to script a repetitive Blender task without writing the Python yourself."],"notFor":["Unsaved work: the execute_blender_code tool runs any Python inside Blender, and the README says to save before using it.","Running two clients at once: only one MCP server instance should connect to Blender at a time.","Official support: it is a community project, not made by the Blender Foundation."],"requirements":["Blender with the MCP for Blender addon enabled","uv (installed by the setup script) and Python 3.10 or newer"],"connect":[{"label":"One-command setup on macOS or Linux (installs uv, configures clients, installs the addon)","command":"curl -LsSf https://www.mcp-for-blender.com/install.sh | sh"},{"label":"Setup when uv is already installed","command":"uvx mcp-for-blender setup"},{"label":"Claude Code, manual","command":"claude mcp add blender uvx mcp-for-blender"},{"label":"Codex, manual","command":"codex mcp add blender -- uvx mcp-for-blender"},{"label":"Install the Blender addon, manual route","command":"uvx mcp-for-blender install-addon"}],"connectNote":"After a manual install, enable Interface: MCP for Blender in Blender under Edit, Preferences, Add-ons. Fully quit and reopen your AI app after setup. In the 3D viewport, press N and open the MCP for Blender tab to check the server is running.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/brave-search","slug":"brave-search","name":"Brave Search","title":"Brave Search MCP server: web, news, image and local search for AI agents","summary":"The Brave Search MCP server gives AI agents Brave's independent search index: web, news, image, video, local and place search, plus pre-extracted page text for grounding and AI summaries.","category":"web-search-and-scraping","publisher":"Brave","publisherUrl":"https://brave.com/search/api/","official":true,"repo":"brave/brave-search-mcp-server","stars":1479,"repoLicense":"MIT","lastPush":"2026-09-30","registryName":"io.github.brave/brave-search-mcp-server","freeTier":"free-plan","cost":"Free and open source under MIT; searches are billed to your Brave Search API account, which has a monthly free credit.","pricingUrl":"https://brave.com/search/api/","clients":["Claude Desktop","VS Code","fx"],"tools":["brave_web_search","brave_local_search","brave_place_search","brave_image_search","brave_video_search","brave_news_search","brave_summarizer","brave_llm_context"],"whatItDoes":["The Brave Search MCP server connects an AI agent to the Brave Search API, which queries Brave's own web index rather than reselling Google or Bing results. The main tool runs a web search with filters for country, language, freshness, safe search and result types, and can apply Goggles, Brave's custom re-ranking rules. Separate tools search news, images, videos, local businesses and points of interest around a location.","Brave's server also has two tools aimed at language models. The LLM context tool returns pre-extracted text from the top results, with limits on tokens, URLs and snippets, so the agent gets readable source material in one call. The summarizer tool turns a web search into an AI-written summary with optional inline references, using a key returned by a search made with summaries turned on.","Brave publishes the server on npm and as a Docker image, and it runs locally over stdio by default or over HTTP when you ask for it. You can allow or block single tools with environment variables. The GitHub repo had 1,478 stars on 2026-10-01; version 2 changed the default transport to stdio and stopped returning base64 image data."],"useItWhen":["You want search results from an index that is not Google's or Bing's.","Your agent needs recent news on a topic, filtered to the last day or week.","You need local businesses or places near a city or coordinates, with hours and ratings.","You want short grounded text from search results without a separate scraping step."],"notFor":["Reading full pages or crawling a site: the server returns results and extracted snippets, not whole pages. Firecrawl covers that.","Use without an API key: every call needs a Brave Search API key, and requests beyond the monthly free credit are billed."],"requirements":["A Brave Search API key","Node.js or Docker to run the server locally"],"connect":[{"label":"Claude Desktop with Docker, in claude_desktop_config.json","command":"{\"mcpServers\": {\"brave-search\": {\"command\": \"docker\", \"args\": [\"run\", \"-i\", \"--rm\", \"-e\", \"BRAVE_API_KEY\", \"docker.io/mcp/brave-search\"], \"env\": {\"BRAVE_API_KEY\": \"YOUR_API_KEY_HERE\"}}}}"},{"label":"VS Code or any client with npx, over stdio","command":"{\"servers\": {\"brave-search-mcp-server\": {\"command\": \"npx\", \"args\": [\"-y\", \"@brave/brave-search-mcp-server\", \"--transport\", \"stdio\"], \"env\": {\"BRAVE_API_KEY\": \"YOUR_API_KEY_HERE\"}}}}"}],"connectNote":"Create the API key in the Brave Search API dashboard. To limit what the agent can call, set BRAVE_MCP_ENABLED_TOOLS or BRAVE_MCP_DISABLED_TOOLS to a space-separated list of tool names.","faq":[{"q":"Is the Brave Search MCP server free?","a":"The server is free and open source. Brave gives every API account a monthly free credit, and requests past it are billed by the request (pricing page, 2026-10-01)."},{"q":"Does Brave Search MCP use Google results?","a":"No. The Brave Search API serves results from Brave's own independent web index."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/browser-use","slug":"browser-use","name":"Browser Use","title":"Browser Use MCP server: an autonomous browser agent for AI tools","summary":"The Browser Use MCP server gives AI agents the Browser Use browsing agent, either as a local open-source server with direct browser tools or as a cloud server that runs whole web tasks.","category":"browser-automation","publisher":"Browser Use","publisherUrl":"https://browser-use.com","official":true,"repo":"browser-use/browser-use","stars":116955,"repoLicense":"MIT","lastPush":"2026-10-01","registryName":"com.browser-use/browser-use","remoteUrl":"https://api.browser-use.com/v3/mcp","freeTier":"free-plan","cost":"The local server is free and open source under MIT and uses your own LLM API key; the cloud server is pay as you go, with free signup credit for eligible accounts.","pricingUrl":"https://browser-use.com/pricing","clients":["Claude Code","Claude Desktop","Cursor","Windsurf"],"tools":["run_session","get_session","send_task","stop_session","list_sessions","list_browser_profiles","browser_navigate","browser_click","browser_type","browser_get_state","browser_extract_content","retry_with_browser_use_agent"],"whatItDoes":["The Browser Use MCP server connects an AI agent to Browser Use, an open-source library in which a language model operates a real Chrome browser to finish a task. It comes in two forms. The cloud server takes a whole task in plain words, such as finding a flight or filling in an application, runs it in Browser Use's cloud browsers, and reports status, cost, a live view URL and the result, optionally shaped by a JSON schema.","The local server runs on your machine through uvx and gives the agent lower-level tools: navigate, click and type by element index, read the page state, scroll, manage tabs, extract content, get HTML and take screenshots. When direct control fails, a fallback tool hands the whole task to the Browser Use agent. Cloud sessions can keep the browser alive for follow-up tasks and use saved browser profiles for sites that need a login.","The browser-use repo on GitHub, which holds the local MCP server, had 116,895 stars on 2026-10-01. Those stars are for the whole library, not for the MCP server alone."],"useItWhen":["You want to hand off a whole web chore, like comparing prices on several shops, and get back the result.","You need a browser task to run in the cloud while your own machine stays free.","You already use the Browser Use library and want the same agent inside your editor.","A task needs a logged-in session that you saved as a browser profile."],"notFor":["Cost control without watching: each run spends model tokens plus browser time, and the local server needs your own LLM API key.","Step-by-step deterministic testing; Playwright MCP gives exact, repeatable actions."],"requirements":["Cloud: a Browser Use Cloud API key","Local: Python with uv, Chrome, and an OpenAI or Anthropic API key"],"connect":[{"label":"Claude Code, cloud server","command":"claude mcp add -t http -H \"x-browser-use-api-key: YOUR_API_KEY\" browser-use https://api.browser-use.com/v3/mcp"},{"label":"Cursor or Claude Desktop, cloud server","command":"{\"mcpServers\": {\"browser-use\": {\"url\": \"https://api.browser-use.com/v3/mcp\", \"headers\": {\"x-browser-use-api-key\": \"YOUR_API_KEY\"}}}}"},{"label":"Claude Code, local open-source server","command":"claude mcp add browser-use -- uvx --from 'browser-use[cli]' browser-use --mcp"}],"connectNote":"Get a cloud API key at cloud.browser-use.com. For the local server, set OPENAI_API_KEY or ANTHROPIC_API_KEY in its environment, and in Claude Desktop use the full path to uvx if the app cannot find it.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/browserbase","slug":"browserbase","name":"Browserbase","title":"Browserbase MCP server: cloud browsers for AI agents","summary":"The Browserbase MCP server gives AI agents a browser that runs in Browserbase's cloud, controlled with plain-language commands through Stagehand: navigate, act on the page, observe elements and extract data.","category":"browser-automation","publisher":"Browserbase","publisherUrl":"https://www.browserbase.com","official":true,"registryName":"io.github.browserbase/mcp-server-browserbase","remoteUrl":"https://mcp.browserbase.com/mcp","freeTier":"free-plan","cost":"Browser time is billed to your Browserbase account; the Free plan has a small monthly allowance, and paid plans add browser hours and concurrency.","pricingUrl":"https://www.browserbase.com/pricing","clients":["Claude Code","Cursor","ChatGPT"],"tools":["start","end","navigate","act","observe","extract"],"whatItDoes":["The Browserbase MCP server connects an AI agent to Browserbase, a service that runs headless browsers in the cloud. It is built on Stagehand, Browserbase's open-source automation framework, so the agent gives instructions in plain language, such as \"click the login button\" or \"fill in the contact form\", instead of writing selectors. A model reads the page and turns each instruction into actions.","Browserbase's server keeps its tool list short: start and end open and close a cloud browser session, navigate goes to a URL, act performs an action, observe finds the elements you can act on, and extract pulls data from the page, optionally guided by an instruction. Every tool takes a session ID, which lets clients such as ChatGPT that reconnect on each call keep working in the same browser. Options turn on proxies, Browserbase Verified for sites with bot checks, and keep-alive sessions.","Browserbase runs the server at mcp.browserbase.com, and its docs recommend that hosted endpoint over running it locally with npx. The open-source repo browserbase/mcp-server-browserbase was archived in 2026 and is kept only as a reference, so the hosted server is the maintained route."],"useItWhen":["You want browser automation without running Chrome on your own machine.","Your agent needs many browsers at once, for example to check prices on dozens of sites.","You prefer to describe actions in words and let the server work out the clicks.","A site blocks ordinary headless browsers and you need proxies or verified browsing."],"notFor":["Free heavy use: the Free plan includes only a little browser time, and every session runs on paid cloud time after that.","Precise, repeatable test steps; plain-language actions go through a model and can vary between runs, where Playwright MCP is deterministic."],"requirements":["A Browserbase account and API key"],"connect":[{"label":"Claude Code","command":"claude mcp add --transport http browserbase \"https://mcp.browserbase.com/mcp?browserbaseApiKey=YOUR_BROWSERBASE_API_KEY\""},{"label":"Clients that support Streamable HTTP","command":"{\"mcpServers\": {\"browserbase\": {\"type\": \"http\", \"url\": \"https://mcp.browserbase.com/mcp\"}}}"}],"connectNote":"The hosted endpoint needs your API key as the browserbaseApiKey query parameter for tool calls. Add modelName and modelApiKey to use a model other than the default google/gemini-2.5-flash-lite; Cursor has a one-click install on Browserbase's setup page.","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/chrome-devtools","slug":"chrome-devtools","name":"Chrome DevTools","title":"Chrome DevTools MCP server: debug and profile Chrome with AI agents","summary":"The Chrome DevTools MCP server gives AI agents control of a live Chrome browser with DevTools access: performance traces, network requests, console errors, Lighthouse audits, memory snapshots and page automation.","category":"browser-automation","publisher":"Google Chrome DevTools team","publisherUrl":"https://developer.chrome.com","official":true,"repo":"ChromeDevTools/chrome-devtools-mcp","stars":52860,"repoLicense":"Apache-2.0","lastPush":"2026-10-01","registryName":"io.github.ChromeDevTools/chrome-devtools-mcp","freeTier":"free","cost":"Free and open source under Apache-2.0; it drives your own Chrome locally, with no account or usage fees.","pricingUrl":"https://github.com/ChromeDevTools/chrome-devtools-mcp","clients":["Claude Code","Codex","Gemini CLI","Gemini Code Assist","GitHub Copilot CLI","VS Code","Visual Studio","Cursor","Windsurf","Cline","Google Antigravity","Amp","Devin CLI","Factory CLI","Grok Build CLI","JetBrains AI Assistant","Junie","Kiro","Mistral Vibe","OpenCode","Qoder","Warp"],"tools":["navigate_page","click","fill_form","take_snapshot","take_screenshot","list_console_messages","list_network_requests","get_network_request","performance_start_trace","performance_analyze_insight","lighthouse_audit","take_heapsnapshot"],"whatItDoes":["The Chrome DevTools MCP server, which Google calls Chrome DevTools for agents, lets an AI coding agent control and inspect a live Chrome browser. Its edge over plain automation is the DevTools side: the agent can record a performance trace and read the insights out of it, list and open network requests, read console messages with source-mapped stack traces, read computed CSS, and run a Lighthouse audit on a page.","Chrome DevTools MCP also covers the usual page actions through Puppeteer, which waits for each action to finish: navigate, click, fill forms, type, drag, upload files, handle dialogs and emulate devices or slow networks. Recent versions add memory work, where the agent takes heap snapshots and compares them to find leaks, plus screencasts and Chrome extension tools. A slim mode keeps only basic browser tools for agents that need fewer tokens.","The server runs locally with npx and starts Chrome the first time a tool needs it; it can also attach to a Chrome you already have open. The GitHub repo had 52,839 stars on 2026-10-01."],"useItWhen":["Your page is slow and you want the agent to trace it and name the cause.","The agent changed code and should confirm there are no new console errors or failed requests.","You want a Lighthouse score and fixes for accessibility or SEO issues.","You suspect a memory leak and want heap snapshots compared before and after an action."],"notFor":["Firefox or Safari: Google officially supports only Chrome and Chrome for Testing.","Private browsing data: the agent sees everything in the browser it controls, and Google collects usage statistics by default (turn off with --no-usage-statistics)."],"requirements":["Node.js LTS and npm","Google Chrome, current stable or newer"],"connect":[{"label":"Claude Code","command":"claude mcp add chrome-devtools --scope user npx chrome-devtools-mcp@latest"},{"label":"Codex","command":"codex mcp add chrome-devtools -- npx chrome-devtools-mcp@latest"},{"label":"Gemini CLI","command":"gemini mcp add chrome-devtools npx chrome-devtools-mcp@latest"},{"label":"Cursor and other clients with a JSON config","command":"{\"mcpServers\": {\"chrome-devtools\": {\"command\": \"npx\", \"args\": [\"-y\", \"chrome-devtools-mcp@latest\"]}}}"}],"connectNote":"Add --slim --headless to the args for basic browsing with fewer tools and no window. Performance tools may send page URLs to Google's CrUX API for real-user data; --no-performance-crux turns that off.","faq":[{"q":"Is the Chrome DevTools MCP server free?","a":"Yes. Google publishes it under Apache-2.0, and it runs locally against your own Chrome."},{"q":"Chrome DevTools MCP or Playwright MCP?","a":"Chrome DevTools MCP is stronger for debugging: performance traces, Lighthouse, network detail and memory snapshots, in Chrome only. Playwright MCP covers more browsers and is built around automation and tests."},{"q":"Does Chrome DevTools MCP send data to Google?","a":"It collects usage statistics such as tool success rates by default, and performance tools can query the CrUX API. Both can be switched off with flags, and statistics stay off when the CI variable is set."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/cloudflare","slug":"cloudflare","name":"Cloudflare","title":"Cloudflare MCP server: Workers, logs, DNS and docs for AI agents","summary":"The Cloudflare MCP servers give AI agents access to a Cloudflare account through hosted endpoints: read Workers logs and builds, manage bindings, check DNS analytics, search the docs and render web pages.","category":"cloud-and-devops","publisher":"Cloudflare","publisherUrl":"https://www.cloudflare.com","official":true,"repo":"cloudflare/mcp-server-cloudflare","stars":4344,"repoLicense":"Apache-2.0","lastPush":"2026-10-01","registryName":"com.cloudflare.mcp/mcp","remoteUrl":"https://mcp.cloudflare.com/mcp","freeTier":"free-plan","cost":"Free to connect and open source under Apache-2.0; some features need a paid Workers plan, and the products the tools use are billed under your Cloudflare plan.","pricingUrl":"https://github.com/cloudflare/mcp-server-cloudflare","clients":["Claude","Cursor","Cloudflare AI Playground","OpenAI Responses API"],"whatItDoes":["Cloudflare runs a set of hosted MCP servers, one per product area, that connect AI agents to a Cloudflare account. The domain servers in this repo cover Workers Bindings for storage, AI and compute, Workers Builds, Observability for logs and analytics, Browser Run for fetching pages as markdown and screenshots, Logpush, AI Gateway logs, AutoRAG, DNS Analytics, Digital Experience Monitoring, Cloudflare One CASB, Radar and the Cloudflare docs.","Cloudflare recommends a separate Code Mode server at mcp.cloudflare.com/mcp, kept in the cloudflare/mcp repo, for broad access across Cloudflare's APIs through a small set of general tools that write and run code. The domain servers fit better when you work in one product and want typed, guided tools. Agents can read configuration, suggest changes, and make those changes in your account.","Every Cloudflare server speaks Streamable HTTP at a /mcp path, with /sse kept only as an alias; legacy SSE clients get an error. Sign-in is OAuth, or an API token with the right scopes when calling from OpenAI's Responses API. The repo had 4,342 GitHub stars on 2026-10-01 and was pushed the same day."],"useItWhen":["A Worker is throwing errors and you want the agent to read its logs and find the cause.","You want the agent to set up KV, R2, D1 or AI bindings for a new Worker.","You need current Cloudflare docs while writing Workers code.","You want a summary of DNS performance or a Workers Builds failure."],"notFor":["Long chained queries in small context windows: Cloudflare warns the Observability server can trigger many tool calls and hit the client's context limit, so keep requests narrow.","Unreviewed changes: the servers can modify account settings and resources, so confirm each write."],"requirements":["A Cloudflare account; a paid Workers plan for some features","An MCP client that supports remote servers, or an API token for API-based clients"],"connect":[{"label":"Code Mode server, broad API access (recommended by Cloudflare)","command":"https://mcp.cloudflare.com/mcp"},{"label":"Documentation server","command":"https://docs.mcp.cloudflare.com/mcp"},{"label":"Observability server, Workers logs and analytics","command":"https://observability.mcp.cloudflare.com/mcp"},{"label":"Workers Bindings server","command":"https://bindings.mcp.cloudflare.com/mcp"}],"connectNote":"Add any of these URLs as a remote server in your client and sign in with Cloudflare when prompted; Cloudflare AI Playground also accepts them. Other servers follow the same pattern, such as builds.mcp.cloudflare.com/mcp and radar.mcp.cloudflare.com/mcp. For OpenAI's Responses API, create an API token in the dashboard with the scopes that server needs.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/context7","slug":"context7","name":"Context7","title":"Context7 MCP server: up-to-date library docs for AI coding agents","summary":"The Context7 MCP server gives AI coding agents current, version-specific documentation and code examples for thousands of libraries, fetched into the prompt so the agent stops guessing at outdated APIs.","category":"docs-and-code-context","publisher":"Upstash","publisherUrl":"https://context7.com","official":true,"repo":"upstash/context7","stars":62586,"repoLicense":"MIT","lastPush":"2026-10-01","registryName":"io.github.upstash/context7","remoteUrl":"https://mcp.context7.com/mcp","freeTier":"free-plan","cost":"The server code is MIT; Context7's Free plan covers public libraries with a monthly call limit, and paid seats add private repositories and more calls.","pricingUrl":"https://context7.com/plans","clients":["Claude Code","Cursor","OpenAI Codex","VS Code","Claude Desktop","ChatGPT","Gemini CLI","GitHub Copilot CLI","Copilot Coding Agent","Google Antigravity","Windsurf","Kiro","Cline","Roo Code","Kilo Code","Zed","JetBrains AI Assistant","Opencode","Amp","Warp","LM Studio","Perplexity Desktop","Amazon Q Developer CLI","Qwen Code","OpenClaw","Hermes"],"tools":["resolve-library-id","query-docs"],"whatItDoes":["The Context7 MCP server connects an AI coding agent to Context7, an index of library and framework documentation run by Upstash. When the agent is about to write code against a library, it first resolves the library name to a Context7 ID such as /vercel/next.js, then asks for the docs and code examples that match the task. Context7 returns the relevant snippets for the version you name, which cuts down on invented APIs and code written for an old release.","Context7 keeps its MCP surface to two tools, which costs little context. You can trigger it by adding \"use context7\" to a prompt, name a library ID directly to skip the lookup, or add a rule to your agent's instructions so it reaches for Context7 on any library question. Library owners can submit their own docs, and Pro teams can index private repositories.","Upstash hosts the server at mcp.context7.com and also publishes it on npm to run locally. Its ctx7 CLI can instead install Context7 as a skill that calls the CLI, with no MCP at all. The GitHub repo had 62,576 stars on 2026-10-01; the backend, parser and crawler are private, and only the MCP server code is open."],"useItWhen":["Your agent keeps writing code for an older version of a framework.","You are working with a library that changed recently or that the model barely knows.","You want short, relevant doc snippets in context instead of pasting whole pages.","Your team wants its own private libraries searchable by the agent."],"notFor":["Guaranteed accuracy: Context7 says library entries are community-contributed and it cannot vouch for every one, so check critical code against the official docs.","Heavy free use: the Free plan has a monthly call limit."],"requirements":["A free Context7 API key or OAuth sign-in for higher rate limits","Node.js 18 or newer for the ctx7 setup command or a local server"],"connect":[{"label":"Guided setup for your agents (OAuth, picks MCP or CLI plus skill)","command":"npx ctx7 setup"},{"label":"Claude Code, remote server","command":"claude mcp add --scope user --header \"Authorization: Bearer YOUR_API_KEY\" --transport http context7 https://mcp.context7.com/mcp"},{"label":"Codex","command":"codex mcp add context7 -- npx -y @upstash/context7-mcp --api-key YOUR_API_KEY"},{"label":"Any client with a JSON config, remote server","command":"{\"mcpServers\": {\"context7\": {\"url\": \"https://mcp.context7.com/mcp\", \"headers\": {\"Authorization\": \"Bearer YOUR_API_KEY\"}}}}"}],"connectNote":"Get a free API key at context7.com/dashboard, or use the OAuth endpoint https://mcp.context7.com/mcp/oauth in clients that support it. Context7's client page has setup steps for more than 40 clients, including ChatGPT through developer mode.","faq":[{"q":"Is the Context7 MCP server free?","a":"Yes for most solo use: the Free plan covers public libraries with a monthly call limit. Paid seats add private repos and more calls (plans page, 2026-10-01)."},{"q":"Does Context7 work with ChatGPT and Codex?","a":"Yes. Context7 documents setup for ChatGPT web and desktop through developer mode apps, a Codex plugin and a codex mcp add command, as well as Cursor, Gemini CLI, Copilot and many others."},{"q":"How do I make the agent use Context7 automatically?","a":"Add a rule to your agent's instructions, for example in CLAUDE.md or Cursor rules, telling it to use Context7 for library docs. The npx ctx7 setup command installs a skill that does this for you."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/dbhub","slug":"dbhub","name":"DBHub","title":"DBHub MCP server: Postgres, MySQL, SQL Server and SQLite for AI agents","summary":"DBHub is an open-source MCP server by Bytebase that gives AI agents SQL access to PostgreSQL, MySQL, MariaDB, SQL Server, Oracle and SQLite through one server with two default tools.","category":"databases","publisher":"Bytebase","publisherUrl":"https://dbhub.ai","official":false,"repo":"bytebase/dbhub","stars":3593,"repoLicense":"MIT","lastPush":"2026-10-01","registryName":"io.github.bytebase/dbhub","freeTier":"free","cost":"Free and open source under MIT and run on your own machine, with no paid tier; Bytebase's enterprise product is separate.","pricingUrl":"https://github.com/bytebase/dbhub","clients":["Claude Code","Claude Desktop","Cursor","VS Code","GitHub Copilot CLI","Codex","Dify","LibreChat"],"tools":["execute_sql","search_objects","explain_sql","health_check"],"whatItDoes":["DBHub is a database MCP server from Bytebase, the company behind a database DevOps platform. One DBHub process connects an AI agent to PostgreSQL, MySQL, MariaDB, SQL Server, Oracle or SQLite, and a TOML file lets it hold several connections at once, for example a local dev database and a read-only replica. It supports SSH tunnels and SSL for databases behind a bastion host.","DBHub keeps its default tool list to two: execute_sql runs queries with transaction support, and search_objects explores schemas, tables, columns, indexes and procedures a step at a time. Bytebase measures this at about 1,400 tokens of context, against roughly 19,000 for some alternatives. Two opt-in tools show a query plan and report connection pool health, and you can define your own parameterized SQL tools in the config file. A built-in web workbench runs the same tools without an MCP client.","Guardrails include a read-only mode, a row limit and a query timeout. DBHub runs through npx (Node.js 22.5 or newer), Docker, an MCP Bundle file or a Claude Code plugin; the bundle and plugin are read-only by design. Claude Code's own docs use DBHub as the example for querying PostgreSQL. The repo had 3,589 GitHub stars on 2026-10-01."],"useItWhen":["You work with more than one database engine and want a single MCP server for all of them.","You want the agent to explore an unfamiliar schema without filling its context with tool definitions.","You want to give non-technical teammates read-only views of data through Claude Desktop or VS Code.","You need to reach a database behind an SSH bastion host."],"notFor":["Deep Postgres tuning such as index recommendations and health reports; Postgres MCP Pro specializes in that.","Write access on production without care: unless read-only mode is on, execute_sql can change data."],"requirements":["Node.js 22.5 or newer, or Docker","A database connection string (DSN)"],"connect":[{"label":"Claude Code, as a plugin (prompts for the connection string and stores it securely)","command":"/plugin marketplace add bytebase/dbhub"},{"label":"Then install the plugin","command":"/plugin install dbhub@dbhub"},{"label":"Codex, trying the built-in demo database first","command":"codex mcp add dbhub -- npx @bytebase/dbhub@latest --transport stdio --demo"},{"label":"Run as an HTTP server for any client (add your own --dsn instead of --demo)","command":"npx @bytebase/dbhub@latest --transport http --port 8080 --demo"}],"connectNote":"Pass your database with --dsn, for example a postgres:// URI with user and password, or keep several connections in a dbhub.toml file and use --config. From Docker, use host.docker.internal instead of localhost to reach a database on your machine.","faq":[{"q":"Is DBHub free?","a":"Yes. DBHub is open source under MIT and runs on your own machine with no paid tier."},{"q":"Which databases does DBHub support?","a":"PostgreSQL, MySQL, MariaDB, SQL Server, Oracle and SQLite (README read 2026-10-01). SQLite needs no extra driver because DBHub uses Node's built-in module."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/deepwiki","slug":"deepwiki","name":"DeepWiki","title":"DeepWiki MCP server: ask questions about any public GitHub repo","summary":"The DeepWiki MCP server gives AI agents DeepWiki's generated documentation for public GitHub repositories, plus an ask tool that answers questions about a repo's code, free and without sign-in.","category":"docs-and-code-context","publisher":"Cognition (makers of Devin)","publisherUrl":"https://deepwiki.com","official":true,"remoteUrl":"https://mcp.deepwiki.com/mcp","freeTier":"free","cost":"Free hosted server with no sign-in, for public repositories only; private repositories need a paid Devin account and a separate server.","pricingUrl":"https://docs.devin.ai/work-with-devin/deepwiki-mcp","clients":["Claude Code","Claude.ai","Cursor","Devin Desktop","OpenAI API"],"tools":["read_wiki_structure","read_wiki_contents","ask_question"],"whatItDoes":["The DeepWiki MCP server connects an AI agent to DeepWiki, a site by Cognition, the company behind the Devin coding agent, that generates wiki-style documentation for public GitHub repositories. With it, an agent can list the documentation topics DeepWiki has written for a repo, read those pages, and ask a free-form question about the repo and get an answer grounded in its code, the same feature DeepWiki's site calls Ask Devin.","DeepWiki's server is useful when an agent has to work with an open-source project it does not know well: how a library structures its plugins, where a function is defined, or how two modules talk to each other. Because the answer comes from DeepWiki's index of the repository rather than from the model's memory, it reflects the code DeepWiki has indexed, not an older training snapshot.","Cognition runs the server at mcp.deepwiki.com with no API key and no account. The /mcp endpoint uses Streamable HTTP and is the recommended one; the older /sse endpoint is being retired. There is no public source repository for the server."],"useItWhen":["You want your agent to understand an unfamiliar open-source library before changing code that uses it.","You need a quick answer to \"how does X work in repo Y\" without cloning the repo.","You want a free docs tool that needs no key and works in any MCP client."],"notFor":["Private code: the free server only covers public repositories, and answers come from DeepWiki's index, which can lag behind the latest commits."],"connect":[{"label":"Claude Code","command":"claude mcp add -s user -t http deepwiki https://mcp.deepwiki.com/mcp"},{"label":"Cursor and most other clients","command":"{\"mcpServers\": {\"deepwiki\": {\"url\": \"https://mcp.deepwiki.com/mcp\"}}}"},{"label":"Devin Desktop","command":"{\"mcpServers\": {\"deepwiki\": {\"serverUrl\": \"https://mcp.deepwiki.com/mcp\"}}}"}],"connectNote":"No sign-in is needed. Clients differ on the URL field name: Devin Desktop uses serverUrl and most others use url, and a wrong field makes the client ignore the server silently.","faq":[{"q":"Is the DeepWiki MCP server free?","a":"Yes. Cognition's docs call it free and say it needs no authentication; it covers public GitHub repositories only (docs read 2026-10-01)."},{"q":"Does DeepWiki MCP work with ChatGPT or the OpenAI API?","a":"Cognition's docs say the /mcp endpoint works with OpenAI and link OpenAI's remote MCP guide, which uses DeepWiki as an example for the OpenAI API. It also works in Claude, Cursor and Devin."},{"q":"Can DeepWiki MCP read my private repo?","a":"No. For private repositories Cognition points to the Devin MCP server, which needs a Devin account and API key."}],"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/exa","slug":"exa","name":"Exa","title":"Exa MCP server: web search and page fetching for AI agents","summary":"The Exa MCP server gives AI agents web search and page reading through Exa's search engine, plus optional advanced filters and Exa Agent runs for multi-step research with structured output.","category":"web-search-and-scraping","publisher":"Exa","publisherUrl":"https://exa.ai","official":true,"repo":"exa-labs/exa-mcp-server","stars":5069,"repoLicense":"MIT","lastPush":"2026-09-30","registryName":"ai.exa/exa","remoteUrl":"https://mcp.exa.ai/mcp","freeTier":"free-plan","cost":"The hosted server works without sign-in at free rate limits; signed-in use is pay as you go, with monthly free credits for new accounts.","pricingUrl":"https://exa.ai/pricing","clients":["ChatGPT","Codex","Claude Code","Claude Desktop","Claude.ai","Cursor","VS Code","Grok Build","Gemini CLI","Windsurf","Zed","Kiro","LM Studio","OpenCode","Google Antigravity","Warp","Replit","v0"],"tools":["web_search_exa","web_fetch_exa","web_search_advanced_exa","agent_run"],"whatItDoes":["The Exa MCP server connects an AI agent to Exa, a search engine built to return results for machines rather than people. By default it adds two tools: one searches the web and returns clean page content for each hit, the other reads one or more URLs and returns their full text as Markdown. That covers the usual need of an agent that has to look something up and then read the source.","Two more tools turn on when you name them in the connection URL. Advanced search adds filters for domains, dates and categories, plus highlights, summaries and subpage crawling. The agent_run tool hands a task to Exa Agent, which does multi-step research, builds lists or enriches records and can return the findings in a JSON shape you define; it needs a signed-in account or API key.","Exa runs the server itself at mcp.exa.ai, so setup is a URL or a plugin rather than a local install. It works anonymously at rate limits, and OAuth sign-in or an API key moves calls onto your own plan. The GitHub repo had 5,070 stars on 2026-10-01 and also ships agent skills that load with the plugin installs."],"useItWhen":["Your coding agent needs current docs, release notes or error reports from the web.","You want search results that already include page text, so the agent skips a fetch step.","You are building a list of companies or people that match criteria and want it as JSON.","You want web search in ChatGPT, Codex, Claude or Grok Build through one provider."],"notFor":["Heavy anonymous use: keyless calls hit Exa's free rate limits quickly (HTTP 429), and agent_run never runs without an account."],"requirements":["Nothing for keyless search; an Exa account or API key for higher limits and Exa Agent"],"connect":[{"label":"Claude Code, as a plugin","command":"claude plugin install exa@claude-plugins-official"},{"label":"Codex","command":"codex mcp add exa --url https://mcp.exa.ai/mcp"},{"label":"Any MCP client with a JSON config","command":"{\"mcpServers\": {\"exa\": {\"type\": \"streamable-http\", \"url\": \"https://mcp.exa.ai/mcp\"}}}"},{"label":"Turn on extra tools by listing them in the URL (this replaces the defaults)","command":"https://mcp.exa.ai/mcp?tools=web_search_exa,web_fetch_exa,agent_run"}],"connectNote":"In ChatGPT and in Claude Desktop or Claude.ai, add Exa from the app's plugin or connector directory and sign in. To force sign-in on any client, connect to https://mcp.exa.ai/mcp?login. An API key from the Exa dashboard can go in an Authorization: Bearer or x-api-key header.","faq":[{"q":"Is the Exa MCP server free?","a":"It can be used free: the hosted endpoint answers without an account at rate limits, and new accounts get monthly free credits. Past that, Exa bills per request (pricing page, 2026-10-01)."},{"q":"Does the Exa MCP server work with ChatGPT?","a":"Yes. Exa publishes a plugin in OpenAI's directory that works in both ChatGPT and Codex, and its docs also cover Claude, Grok Build, Cursor, Gemini CLI and a dozen other clients."},{"q":"What is the difference between Exa and Firecrawl MCP servers?","a":"Exa is mainly a search engine that returns page text with results. Firecrawl is mainly a scraper and crawler for sites you already know, with page interaction. Many agents use one for finding and the other for extracting."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/figma","slug":"figma","name":"Figma","title":"Figma MCP server: design context and canvas edits for AI agents","summary":"The Figma MCP server gives AI agents design context from Figma files, such as layout, variables and components, to generate code from frames, and can write native content back to the canvas.","category":"design","publisher":"Figma","publisherUrl":"https://www.figma.com","official":true,"repo":"figma/mcp-server-guide","stars":2041,"repoLicense":"None (no license on GitHub)","lastPush":"2026-10-01","registryName":"com.figma.mcp/mcp","remoteUrl":"https://mcp.figma.com/mcp","freeTier":"free-plan","cost":"No separate charge; your tool-call limit depends on your Figma plan and seat, the free Starter plan has a small monthly limit, and canvas writes are free during the beta but will become a paid feature.","pricingUrl":"https://developers.figma.com/docs/figma-mcp-server/rate-limits-access/","clients":["Claude Code","Codex","Cursor","VS Code","Xcode","Gemini CLI"],"tools":["get_design_context","get_variable_defs"],"whatItDoes":["The Figma MCP server is Figma's own server for giving AI coding agents structured design data. Paste a link to a frame or layer and the agent can pull its layout, components and variables, then write code from it. get_design_context returns a React and Tailwind representation of the selection that the agent can translate into any framework, and get_variable_defs extracts the color, spacing and typography tokens used.","The Figma MCP server also works in the other direction. On the remote server, agents can create and modify native Figma content, such as frames, components, variables and auto layout, using your design system as the source. Code Connect links Figma components to components in your codebase so generated code reuses them instead of inventing new ones. Capturing a web page into a Figma design was still rolling out on 2026-10-01.","Figma hosts the server at mcp.figma.com/mcp, and a local desktop server runs inside the Figma app. The GitHub repo holds setup docs and skills, not server code, and had 2,041 stars on 2026-10-02; GitHub shows no license for it. Figma's developer docs state that only clients listed in its MCP catalog can connect."],"useItWhen":["You want the agent to build a component from a specific Figma frame.","You want generated code to use your design tokens instead of hard-coded values.","You want the agent to reuse your existing React components through Code Connect.","You want the agent to draft screens or update components directly in a Figma file."],"notFor":["Free or view-only users who need regular use: the Starter plan and View or Collab seats have low monthly tool-call limits.","Any MCP client you like: Figma only lets clients listed in its MCP catalog connect.","Pixel-perfect output without a tidy file: results depend on components, variables, auto layout and named layers in the design."],"requirements":["A Figma account; a Dev or Full seat on a paid plan for regular use","GitHub Copilot enabled, if you use VS Code"],"connect":[{"label":"Claude Code, plugin with MCP settings and skills (recommended by Figma)","command":"claude plugin install figma@claude-plugins-official"},{"label":"Claude Code, server only","command":"claude mcp add --transport http figma https://mcp.figma.com/mcp"},{"label":"Codex CLI","command":"codex mcp add figma --url https://mcp.figma.com/mcp"},{"label":"Cursor, plugin (type in agent chat)","command":"/add-plugin figma"},{"label":"Gemini CLI extension (then run /mcp auth figma inside gemini)","command":"gemini extensions install https://github.com/figma/mcp-server-guide"},{"label":"VS Code mcp.json","command":"{\"servers\": {\"figma\": {\"type\": \"http\", \"url\": \"https://mcp.figma.com/mcp\"}}}"}],"connectNote":"After adding the server, sign in to Figma when the client prompts you. Then copy a link to a frame or layer and ask the agent to implement it; the agent reads the node ID from the link.","faq":[{"q":"Is the Figma MCP server free?","a":"It has no separate price, but the number of tool calls depends on your plan and seat: the Starter plan and View or Collab seats get only a few calls a month, and Dev or Full seats on paid plans get daily limits (Figma developer docs read 2026-10-02)."},{"q":"Can the Figma MCP server edit designs?","a":"Yes, on the remote server. It can create and change frames, components, variables and auto layout. Figma says this write feature is free during the beta and will later become usage-based and paid."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/firecrawl","slug":"firecrawl","name":"Firecrawl","title":"Firecrawl MCP server: web scraping, crawling and search for AI agents","summary":"The Firecrawl MCP server gives AI agents live web access through Firecrawl: search the web, scrape a URL into Markdown or JSON, map and crawl whole sites, and click through pages.","category":"web-search-and-scraping","publisher":"Firecrawl","publisherUrl":"https://www.firecrawl.dev","official":true,"repo":"firecrawl/firecrawl-mcp-server","stars":7538,"repoLicense":"MIT","lastPush":"2026-10-02","registryName":"io.github.firecrawl/firecrawl-mcp-server","remoteUrl":"https://mcp.firecrawl.dev/v2/mcp","freeTier":"free-plan","cost":"Free and open source under MIT, with a keyless hosted tier that has daily limits; full use spends Firecrawl credits from a free or paid plan.","pricingUrl":"https://www.firecrawl.dev/pricing","clients":["Claude Code","Codex","Cursor","VS Code","Windsurf","OpenCode","Claude Desktop","ChatGPT","Claude.ai"],"tools":["firecrawl_scrape","firecrawl_search","firecrawl_map","firecrawl_crawl","firecrawl_check_crawl_status","firecrawl_parse","firecrawl_agent","firecrawl_agent_status","firecrawl_interact","firecrawl_interact_stop","firecrawl_developer_search","firecrawl_credit_usage"],"whatItDoes":["The Firecrawl MCP server connects an AI agent to Firecrawl, a web data API. With it the agent can run a web search and get ranked results, scrape a known URL into clean Markdown or into JSON that matches a schema you give it, list every URL on a site without fetching them, and crawl a section of a site within limits you set. It also parses files such as PDFs into text.","Firecrawl's server goes past plain fetching. An interact tool clicks, types and navigates on a page before reading it, an agent tool runs multi-source research and returns structured data when you do not know which sites hold the answer, and monitor tools re-check a page on a schedule and report what changed. A developer search tool looks across GitHub issues, pull requests, READMEs and docs for programming questions.","Firecrawl runs the server for you at mcp.firecrawl.dev, so most people connect a URL instead of installing anything. The same open-source code runs locally with npx over stdio or Streamable HTTP, and it can point at a self-hosted Firecrawl instance. The repo had 7,538 GitHub stars on 2026-10-01."],"useItWhen":["Your agent needs the current text of a web page, not what the model remembers.","You want product prices, listings or contacts pulled from a site into JSON.","You need every docs page of a site crawled and fed to the agent as Markdown.","A page only shows its content after a click or a login step.","You want to try web scraping in an agent with no account at all, through the keyless endpoint."],"notFor":["Heavy use: crawl, map and agent need an API key or sign-in, and every call spends credits from a monthly allowance.","Driving a browser step by step across a long session: each interact call runs one turn and returns. Playwright or Chrome DevTools fit that job better."],"requirements":["A Firecrawl account and API key for the full tool set (keyless covers search, scrape and parse)","Node.js, only when running the server locally with npx"],"connect":[{"label":"Claude Code, signing in with your Firecrawl account (finish sign-in with /mcp)","command":"claude mcp add --transport http firecrawl https://mcp.firecrawl.dev/v2/mcp-oauth"},{"label":"Codex, signing in with your Firecrawl account","command":"codex mcp add firecrawl --url https://mcp.firecrawl.dev/v2/mcp-oauth"},{"label":"Cursor or any client with a JSON config, keyless","command":"{\"mcpServers\": {\"firecrawl\": {\"url\": \"https://mcp.firecrawl.dev/v2/mcp\"}}}"},{"label":"Run the open-source server locally with an API key","command":"env FIRECRAWL_API_KEY=fc-YOUR_API_KEY npx -y firecrawl-mcp"}],"connectNote":"In ChatGPT and Claude.ai, Firecrawl points to its native connector in each app's connector directory instead of a URL. With an API key on the hosted endpoint, send it as an Authorization: Bearer header in the client config, never in the URL.","faq":[{"q":"Is the Firecrawl MCP server free?","a":"The server code is free under MIT, and the hosted keyless endpoint lets an agent search, scrape and parse without an account, with daily limits. Beyond that you spend Firecrawl credits from the monthly allowance of the Free plan or a paid plan (pricing page, 2026-10-01)."},{"q":"Does the Firecrawl MCP server work with ChatGPT?","a":"Yes. Firecrawl's docs send ChatGPT users to a native Firecrawl connector in ChatGPT, and give setup lines for Codex, Claude Code, Cursor, OpenCode and any client that reads a JSON MCP config."},{"q":"Can I use it with a self-hosted Firecrawl?","a":"Yes. Run the server locally and set FIRECRAWL_API_URL to your own instance; the API key then becomes optional."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/github","slug":"github","name":"GitHub","title":"GitHub MCP server: repos, issues, pull requests and Actions for AI agents","summary":"The GitHub MCP server gives AI agents access to GitHub through your account: read code and commits, create and review issues and pull requests, check Actions runs and read security alerts.","category":"git-and-code-hosting","publisher":"GitHub","publisherUrl":"https://github.com","official":true,"repo":"github/github-mcp-server","stars":33315,"repoLicense":"MIT","lastPush":"2026-10-01","registryName":"io.github.github/github-mcp-server","remoteUrl":"https://api.githubcopilot.com/mcp/","freeTier":"free","cost":"Free and open source under MIT with no separate charge; it works through your GitHub account and GitHub's normal API rate limits.","pricingUrl":"https://github.com/github/github-mcp-server","clients":["VS Code","Visual Studio","JetBrains IDEs","Xcode","Eclipse","GitHub Copilot CLI","Copilot coding agent","Claude Code","Claude Desktop","Codex","Cursor","Windsurf","Gemini CLI","OpenCode","Zed","Cline","Rovo Dev CLI"],"whatItDoes":["The GitHub MCP server is GitHub's own server for connecting AI agents to GitHub. Through it an agent can browse and search code in any repository you can access, read commits and branches, open, update and comment on issues, create and review pull requests, check GitHub Actions runs and their logs, and read code scanning, secret scanning and Dependabot alerts. It also covers discussions, gists, notifications, projects, labels and organization data.","GitHub's server groups its tools into toolsets, such as repos, issues, pull_requests and actions, and turns on a default set unless you choose others. A read-only mode removes every tool that can change something, and a lockdown mode hides content from public-repo authors without push access to cut the risk of prompt injection through issues and comments. The hosted server adds Copilot tools, such as assigning an issue to the Copilot coding agent, and a GitHub support docs search.","GitHub hosts the server at api.githubcopilot.com/mcp with OAuth or a personal access token, and the same Go code runs locally as a Docker image or binary, which also works with GitHub Enterprise Server. The repo had 33,312 GitHub stars on 2026-10-01."],"useItWhen":["You want the agent to open a pull request, then read the review comments and fix them.","A CI run failed and you want the agent to read the Actions log and find the cause.","You want to triage a backlog of issues: label, comment and close duplicates.","You need a summary of open Dependabot or code scanning alerts across a repo."],"notFor":["Hands-off use with a broad token: the server can push code, merge and close issues with whatever rights your token has, so start in read-only mode or with a fine-grained token.","Reading untrusted public issues without care: text in issues and comments can carry prompt injection; lockdown mode reduces but does not remove the risk."],"requirements":["A GitHub account (OAuth) or a personal access token","Docker or the server binary, only for the local server"],"connect":[{"label":"Claude Code, hosted server with a personal access token","command":"claude mcp add --transport http github https://api.githubcopilot.com/mcp -H \"Authorization: Bearer YOUR_GITHUB_PAT\""},{"label":"Codex, hosted server (token read from an environment variable)","command":"codex mcp add github --url https://api.githubcopilot.com/mcp/ --bearer-token-env-var GITHUB_PAT_TOKEN"},{"label":"VS Code 1.101 or later, hosted server with OAuth","command":"{\"servers\": {\"github\": {\"type\": \"http\", \"url\": \"https://api.githubcopilot.com/mcp/\"}}}"},{"label":"Local server in Docker, read-only","command":"docker run -i --rm -e GITHUB_PERSONAL_ACCESS_TOKEN=<your-token> -e GITHUB_READ_ONLY=1 ghcr.io/github/github-mcp-server"}],"connectNote":"OAuth on the hosted server needs a client that has registered a GitHub App or OAuth App; any client that supports remote servers can use a personal access token instead. Add /readonly to a toolset URL on the hosted server for read-only access.","faq":[{"q":"Is the GitHub MCP server free?","a":"Yes. The server is open source under MIT and GitHub lists no charge for it; calls count against your account's normal GitHub API rate limits (README read 2026-10-01)."},{"q":"Can the GitHub MCP server write, or only read?","a":"It can write: create branches, push files, open and merge pull requests, and edit issues, within your token's permissions. Run it with --read-only, GITHUB_READ_ONLY=1, or a /readonly URL to allow reading only."},{"q":"Does the GitHub MCP server need GitHub Copilot?","a":"No. It works in Claude, Codex, Cursor, Windsurf, Gemini CLI and other non-Copilot clients with a GitHub token. Copilot policies only govern GitHub's own Copilot editors."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/gitlab","slug":"gitlab","name":"GitLab","title":"GitLab MCP server: merge requests, issues and pipelines for AI agents","summary":"The GitLab MCP server gives AI agents access to a GitLab instance through your account: read and create issues and merge requests, commit files, review diffs, and check CI pipelines and jobs.","category":"git-and-code-hosting","publisher":"GitLab","publisherUrl":"https://about.gitlab.com","official":true,"registryName":"com.gitlab/mcp","remoteUrl":"https://gitlab.com/api/v4/mcp","freeTier":"free-plan","cost":"Included with no separate charge in every GitLab tier, Free, Premium and Ultimate; the Duo agent tools need GitLab Duo.","pricingUrl":"https://docs.gitlab.com/user/model_context_protocol/mcp_server/","clients":["Claude Code","Claude Desktop","Cursor","Codex","GitHub Copilot in VS Code","Gemini CLI","Gemini Code Assist","Amazon Q Developer","Kiro","OpenCode","Zed"],"tools":["get_project","get_issue","create_issue","list_merge_requests","get_merge_request_diffs","save_merge_request","save_merge_request_review","add_commit","get_repository_file","get_pipeline_jobs","list_work_items","semantic_search"],"whatItDoes":["The GitLab MCP server is built into GitLab itself and connects AI agents to your GitLab projects. Through it an agent can read projects, files, branches and commits, create and update issues and work items, open, review and accept merge requests, read merge request diffs, notes and conflicts, and commit file changes. On the CI side it lists pipelines, reads jobs and artifacts, and can start or manage a pipeline.","GitLab's server groups its tools into toolsets. The defaults cover core project data, merge requests, work items, the repository and CI; wikis, code security (vulnerabilities) and GitLab Duo agent sessions are opt-in. It also offers search, including semantic code search. Each AI tool registers itself with your GitLab instance through OAuth on first connect, so users sign in with their own GitLab account and see only what they already can.","There is nothing to install: the server lives at /api/v4/mcp on GitLab.com or your own instance. GitLab marks it as beta. An administrator or top-level group owner must first allow access to the MCP server in settings."],"useItWhen":["Your code is on GitLab and you want the agent to open a merge request for its change.","A pipeline failed and you want the agent to read the job log and suggest a fix.","You want to review merge request diffs and leave notes from your editor.","You run self-managed GitLab and need an MCP server that stays inside your instance."],"notFor":["Production-critical automation today: GitLab still labels the server beta, and it can commit, merge and run pipelines with your permissions, so watch what the agent does."],"requirements":["A GitLab account on GitLab.com, Self-Managed or Dedicated","MCP server access allowed for the top-level group (GitLab.com) or the instance","Node.js 20 or later only for clients that connect through mcp-remote"],"connect":[{"label":"Claude Code (replace the host with yours, e.g. gitlab.com)","command":"claude mcp add -s user --transport http GitLab https://<gitlab.example.com>/api/v4/mcp"},{"label":"Codex (then run codex mcp login GitLab)","command":"codex mcp add GitLab --url \"https://<gitlab.example.com>/api/v4/mcp\""},{"label":"Cursor and other clients with HTTP support","command":"{\"mcpServers\": {\"GitLab\": {\"type\": \"http\", \"url\": \"https://<gitlab.example.com>/api/v4/mcp\"}}}"},{"label":"Claude Desktop and other stdio-only clients, through mcp-remote","command":"{\"mcpServers\": {\"GitLab\": {\"command\": \"npx\", \"args\": [\"-y\", \"mcp-remote\", \"https://<gitlab.example.com>/api/v4/mcp\"]}}}"}],"connectNote":"The first connection opens an OAuth sign-in to your GitLab instance. To choose toolsets, send the X-Gitlab-Enabled-Mcp-Server-Toolsets header (behind the mcp_toolsets feature flag since GitLab 19.5).","faq":[{"q":"Is the GitLab MCP server free?","a":"Yes. Since GitLab 19.2 it is part of the Free tier as well as Premium and Ultimate, with no separate charge (GitLab docs, read 2026-10-01)."},{"q":"Is there an official GitLab MCP server?","a":"Yes. GitLab ships its own MCP server inside the product at /api/v4/mcp, in beta since GitLab 18.6. Several community servers exist too, but this one is GitLab's."},{"q":"Does it work with self-managed GitLab?","a":"Yes. It runs on GitLab.com, GitLab Self-Managed and GitLab Dedicated; an administrator must allow access to the MCP server for the instance."}],"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/godot","slug":"godot","name":"Godot","title":"Godot MCP server: run, debug and edit Godot projects with AI agents","summary":"The Godot MCP server gives AI agents control of the Godot engine: launch the editor, run projects in debug mode, capture errors and console output, and create scenes and nodes.","category":"game-and-3d","publisher":"Coding-Solo","publisherUrl":"https://github.com/Coding-Solo","official":false,"repo":"Coding-Solo/godot-mcp","stars":5906,"repoLicense":"MIT","lastPush":"2026-04-16","freeTier":"free","cost":"Free and open source under MIT, and Godot itself is free, so there is nothing to pay.","pricingUrl":"https://github.com/Coding-Solo/godot-mcp","clients":["Claude Code","Cline","Cursor"],"tools":["launch_editor","run_project","get_debug_output","stop_project","get_godot_version","list_projects","get_project_info","create_scene","add_node","load_sprite","save_scene","update_project_uids"],"whatItDoes":["Godot MCP is a community MCP server, written in TypeScript, that connects AI agents to the Godot game engine. Its main use is a feedback loop: the agent writes GDScript, runs the project in debug mode, reads the console output and errors, and fixes what broke, instead of guessing whether its code works.","Godot MCP also edits projects. It can create a scene with a chosen root node type, add nodes with properties, load textures into Sprite2D nodes, export 3D scenes as MeshLibrary resources for GridMap, and save scenes or variants. Simple actions call Godot's command line directly; complex ones go through a single bundled GDScript file that takes the operation as JSON. For Godot 4.4 and later it can read and refresh resource UIDs.","Godot MCP runs with npx and finds the Godot executable on its own, or uses the path in GODOT_PATH. The repo had 5,900 GitHub stars on 2026-10-01, but its last push was on 2026-04-16, more than five months earlier, so check the issue tracker before relying on it with a new Godot release."],"useItWhen":["You want the agent to run your game and fix the errors it sees in the output.","You want a new scene scaffolded with the right node tree.","You want to know the structure of an unfamiliar Godot project.","You need resource UIDs updated after moving files in Godot 4.4 or later."],"notFor":["Projects that need an actively maintained tool: the last push was on 2026-04-16, and newer community Godot servers are listed in the MCP Registry.","Editing scenes while the editor runs: it changes scene files through a script and Godot's command line, not through a live editor plugin."],"requirements":["Godot Engine installed","Node.js 18 or newer and npm"],"connect":[{"label":"Claude Code","command":"claude mcp add godot -- npx @coding-solo/godot-mcp"},{"label":"Claude Code with a custom Godot path and debug logs","command":"claude mcp add godot -e GODOT_PATH=/path/to/godot -e DEBUG=true -- npx @coding-solo/godot-mcp"},{"label":"Cursor .cursor/mcp.json, Cline or any client with a JSON config","command":"{\"mcpServers\": {\"godot\": {\"command\": \"npx\", \"args\": [\"@coding-solo/godot-mcp\"], \"env\": {\"GODOT_PATH\": \"/path/to/godot\", \"DEBUG\": \"true\"}}}}"}],"connectNote":"Restart the client after adding the server. If the server cannot find Godot, set GODOT_PATH to the Godot executable. Project paths must point to a folder containing project.godot.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/jina","slug":"jina","name":"Jina AI","title":"Jina AI MCP server: read URLs, web and arXiv search for AI agents","summary":"The Jina AI MCP server gives AI agents Jina's Reader, Search, Embeddings and Reranker APIs: read any web page or PDF as Markdown, search the web, arXiv and SSRN, and rerank results.","category":"web-search-and-scraping","publisher":"Jina AI (part of Elastic)","publisherUrl":"https://jina.ai","official":true,"repo":"jina-ai/MCP","stars":869,"repoLicense":"Apache-2.0","lastPush":"2026-09-18","remoteUrl":"https://mcp.jina.ai/v1","freeTier":"free-plan","cost":"Free and open source under Apache-2.0; reading pages works without a key at a rate limit, new API keys come with free tokens, and use past that is billed by token.","pricingUrl":"https://jina.ai/reader","clients":["Claude Code","Codex","Claude Desktop","Cursor","LM Studio"],"tools":["primer","read_url","capture_screenshot_url","guess_datetime_url","search_web","search_arxiv","search_ssrn","search_images","search_jina_blog","sort_by_relevance","deduplicate_strings","extract_pdf"],"whatItDoes":["The Jina AI MCP server is a hosted server that connects an AI agent to four Jina APIs. The best-known tool reads a web page or PDF through Jina Reader and returns it as Markdown, or returns only the passages that answer a question you pass in. Other tools capture a screenshot of a page and estimate when a page was published or last updated.","Jina's search tools cover the open web, web images, arXiv preprints and SSRN papers, which makes the server useful for academic lookups as well as general ones. A PDF tool pulls figures, tables and equations out of papers. Two ranking tools use Jina's Reranker and Embeddings models to sort documents by relevance to a query and to drop near-duplicate strings.","The server runs at mcp.jina.ai and has no local install. You can cut the tool list down with query parameters on the URL, such as include_tags=search,read, which saves context tokens in the agent. Jina AI is now part of Elastic, and the GitHub repo had 869 stars on 2026-10-01."],"useItWhen":["Your agent needs a web page or PDF as clean Markdown, quickly and with no setup.","You want to search arXiv or SSRN papers from inside an agent.","You have a long list of search hits and want them reranked by relevance before reading.","You need to know roughly when a page was published."],"notFor":["Search without signing up: search_web, search_arxiv and the ranking tools need a Jina API key, and without one page reading runs at a lower rate limit."],"requirements":["A Jina API key for search, reranking and PDF extraction (reading works without one, at a lower rate limit)"],"connect":[{"label":"Claude Code","command":"claude mcp add -s user --transport http jina https://mcp.jina.ai/v1 --header \"Authorization: Bearer ${JINA_API_KEY}\""},{"label":"Any client that supports remote servers","command":"{\"mcpServers\": {\"jina-mcp-server\": {\"url\": \"https://mcp.jina.ai/v1\", \"headers\": {\"Authorization\": \"Bearer ${JINA_API_KEY}\"}}}}"},{"label":"Codex, in ~/.codex/config.toml","command":"[mcp_servers.jina-mcp-server]\ncommand = \"npx\"\nargs = [\"-y\", \"mcp-remote\", \"https://mcp.jina.ai/v1\", \"--header\", \"Authorization: Bearer ${JINA_API_KEY}\"]"},{"label":"Load only the search and read tools","command":"https://mcp.jina.ai/v1?include_tags=search,read"}],"connectNote":"Some clients do not expand ${JINA_API_KEY}; then paste the real key (it starts with jina_). Clients without remote server support connect through the mcp-remote proxy, as in the Codex example.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/kubernetes","slug":"kubernetes","name":"Kubernetes","title":"Kubernetes MCP server: pods, resources and Helm for AI agents","summary":"The Kubernetes MCP server gives AI agents direct access to Kubernetes and OpenShift clusters: list and edit any resource, read pod logs, exec into pods, view events and install Helm charts.","category":"cloud-and-devops","publisher":"containers (GitHub organization)","publisherUrl":"https://github.com/containers","official":false,"repo":"containers/kubernetes-mcp-server","stars":2140,"repoLicense":"Apache-2.0","lastPush":"2026-10-01","registryName":"io.github.containers/kubernetes-mcp-server","freeTier":"free","cost":"Free and open source under Apache-2.0, with no hosted version or paid tier; you run it against your own clusters.","pricingUrl":"https://github.com/containers/kubernetes-mcp-server","clients":["Claude Code","Claude Desktop","VS Code","VS Code Insiders","Cursor","Goose CLI"],"whatItDoes":["kubernetes-mcp-server is a community MCP server from the containers GitHub organization that connects AI agents to Kubernetes and OpenShift clusters. Through it an agent can create, read, update and delete any resource, list pods across namespaces, read pod logs and resource usage, exec into a pod, run a container image, view events, and install, list or uninstall Helm releases.","kubernetes-mcp-server is written in Go and talks to the Kubernetes API directly instead of wrapping kubectl or helm, so it needs no other tools installed. It ships as a single binary for Linux, macOS and Windows, an npm package, a Python package and a container image. It reads your kubeconfig, follows changes to it, and can work with several clusters at once.","Tools come in toolsets: core and config are on by default, while helm, kcp, kiali, kubevirt, netobserv and tekton are optional. A TOML config sets read_only = true and can deny specific resource types. The project's own Claude Code guide uses read-only mode with a dedicated ServiceAccount. The repo had 2,139 GitHub stars on 2026-10-01, more than the runner-up, Flux159/mcp-server-kubernetes, with 1,594."],"useItWhen":["A pod keeps crashing and you want the agent to read its logs and events.","You want a quick inventory of what runs in a namespace.","You want the agent to install or remove a Helm release in a dev cluster.","You want to troubleshoot a failing Tekton PipelineRun."],"notFor":["Production clusters with an admin kubeconfig: the server can delete resources and exec into pods, so use read_only = true and a limited ServiceAccount.","An official server: no Kubernetes or CNCF project publishes one. The runner-up community option is Flux159/mcp-server-kubernetes (1,594 GitHub stars on 2026-10-01, MIT), which wraps kubectl and helm."],"requirements":["Access to a Kubernetes or OpenShift cluster and a kubeconfig","Node.js for the npx route, or the native binary"],"connect":[{"label":"Claude Desktop or any client with a JSON config","command":"{\"mcpServers\": {\"kubernetes\": {\"command\": \"npx\", \"args\": [\"-y\", \"kubernetes-mcp-server@latest\"]}}}"},{"label":"VS Code","command":"code --add-mcp '{\"name\":\"kubernetes\",\"command\":\"npx\",\"args\":[\"kubernetes-mcp-server@latest\"]}'"},{"label":"Claude Code, read-only with a dedicated kubeconfig (from the project's guide)","command":"claude mcp add-json kubernetes-mcp-server '{\"command\":\"npx\",\"args\":[\"-y\",\"kubernetes-mcp-server@latest\",\"--config\",\"'\"${HOME}\"'/.config/kubernetes-mcp-server.toml\"],\"env\":{\"KUBECONFIG\":\"'${HOME}'/.kube/mcp-viewer.kubeconfig\"}}' -s user"}],"connectNote":"For the Claude Code route, first create ~/.config/kubernetes-mcp-server.toml containing read_only = true, and a ServiceAccount kubeconfig as the project's Kubernetes setup guide describes. Runtime flags such as --read-only are no longer accepted; settings go in the TOML file.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/linear","slug":"linear","name":"Linear","title":"Linear MCP server: issues, projects and comments for AI agents","summary":"The Linear MCP server gives AI agents access to your Linear workspace through Linear's hosted endpoint: find, create and update issues, projects and comments, with a read-only URL when you only need lookups.","category":"productivity","publisher":"Linear","publisherUrl":"https://linear.app","official":true,"registryName":"app.linear/linear","remoteUrl":"https://mcp.linear.app/mcp","freeTier":"free-plan","cost":"Linear's docs list no separate charge or plan restriction for the server; Linear has a Free plan and paid per-user plans.","pricingUrl":"https://linear.app/pricing","clients":["Claude Code","Claude.ai","Claude Desktop","Codex","Cursor","VS Code","Windsurf","Zed","Jules","v0 by Vercel"],"whatItDoes":["The Linear MCP server is Linear's own hosted server for connecting AI agents to a Linear workspace. Once signed in, an agent can look up issues, projects and comments, create new issues, change status, assignee or labels, and post comments, all with the permissions of the account that authorized it. Linear publishes no source code for it: it is a remote service only.","Linear serves the MCP server at two URLs. The default one at mcp.linear.app/mcp can read and write; the one ending in /mcp/readonly allows lookups only, which suits agents that should report on work but never change it. Sign-in is OAuth 2.1 in an interactive flow, and Linear also accepts a Linear API key or bearer token in the Authorization header for clients and automations that cannot open a browser.","Linear documents setup for Claude Code, Codex, Cursor, Claude Desktop and Claude.ai connectors, VS Code, Windsurf, Zed, Jules and v0, and gives the mcp-remote proxy command for any other client that only runs local servers. The deprecated SSE endpoint at /sse still exists as a fallback. The server is listed in the official MCP Registry as app.linear/linear."],"useItWhen":["You want the coding agent to read the Linear issue it is fixing and move it to Done.","You want to file a bug report from an error the agent just found.","You need a summary of a project's open issues or a cycle's progress.","You want a read-only reporting agent that can never edit issues."],"notFor":["Agents you do not watch on the default URL: it can create and edit issues as you, so use the /mcp/readonly URL when lookups are enough.","Self-hosting or auditing the code: Linear publishes no source, so you cannot run or inspect the server yourself."],"requirements":["A Linear account in the workspace you want to connect","Node.js only for clients that connect through the mcp-remote proxy"],"connect":[{"label":"Claude Code","command":"claude mcp add --transport http linear-server https://mcp.linear.app/mcp"},{"label":"Codex (then run: codex mcp login linear)","command":"codex mcp add linear --url https://mcp.linear.app/mcp"},{"label":"VS Code or Windsurf, through mcp-remote","command":"{\"mcpServers\": {\"linear\": {\"command\": \"npx\", \"args\": [\"-y\", \"mcp-remote\", \"https://mcp.linear.app/mcp\"]}}}"},{"label":"Read-only server URL","command":"https://mcp.linear.app/mcp/readonly"}],"connectNote":"Cursor installs Linear from its MCP tools directory, and Claude Desktop and Claude.ai add it as a connector in Settings. Jules uses a Linear API key pasted into its MCP settings.","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/microsoft-learn","slug":"microsoft-learn","name":"Microsoft Learn","title":"Microsoft Learn MCP server: official Microsoft and Azure docs for AI agents","summary":"The Microsoft Learn MCP server gives AI agents search and full-page reading over Microsoft's official documentation, plus official code samples for Azure, .NET and other Microsoft products, free and keyless.","category":"docs-and-code-context","publisher":"Microsoft","publisherUrl":"https://learn.microsoft.com","official":true,"repo":"MicrosoftDocs/mcp","stars":1923,"repoLicense":"CC-BY-4.0","lastPush":"2026-09-29","registryName":"com.microsoft/microsoft-learn-mcp","remoteUrl":"https://learn.microsoft.com/api/mcp","freeTier":"free","cost":"Free; Microsoft runs the server with no API key, login or sign-up.","pricingUrl":"https://github.com/MicrosoftDocs/mcp","clients":["VS Code","Visual Studio","GitHub Copilot CLI","GitHub Copilot coding agent","Claude Code","Claude Desktop","Codex","ChatGPT","Cursor","Windsurf","Gemini CLI","Qwen Code","Cline","Roo Code","Kiro"],"tools":["microsoft_docs_search","microsoft_docs_fetch","microsoft_code_sample_search"],"whatItDoes":["The Microsoft Learn MCP server connects an AI agent to Microsoft Learn, the home of Microsoft's official technical documentation. It has three tools: a semantic search over the docs, a fetch tool that turns any Learn page into Markdown so the agent can read it in full, and a code sample search that finds official snippets for Azure, .NET and other Microsoft technologies, with an optional language filter.","Microsoft built the server to stop agents from inventing SDK methods or Azure CLI flags. Because it searches only first-party Microsoft docs, answers about service limits, region availability or API signatures come from the current documentation rather than from training data or random blogs. A token budget parameter on the URL caps how much text search returns, and an experimental endpoint follows OpenAI's spec for deep research models.","Microsoft hosts the server at learn.microsoft.com/api/mcp; Visual Studio 2022 and 2026 ship with it built in. The same tools exist as a CLI, @microsoft/learn-cli, and as agent skills in the GitHub repo, which had 1,922 stars on 2026-10-01. The repo's license, CC-BY-4.0, covers its documentation content; the server itself is a hosted Microsoft service."],"useItWhen":["You write code against Azure SDKs or .NET and want the agent to check the current API.","You need exact Azure CLI or PowerShell commands for a task.","You want to know a service's limits, quotas or region availability from the official source.","You want a free docs server that needs no key in any MCP client."],"notFor":["Anything outside Microsoft's documentation: it does not search the open web or other vendors' docs; pair it with Context7 for third-party libraries."],"connect":[{"label":"Claude Code, as a plugin with the MCP server and skills","command":"/plugin install microsoft-docs@claude-plugins-official"},{"label":"Codex","command":"codex mcp add \"microsoft-learn\" --url \"https://learn.microsoft.com/api/mcp\""},{"label":"GitHub Copilot CLI","command":"/plugin install microsoftdocs/mcp"},{"label":"Standard config for VS Code and most clients","command":"{\"servers\": {\"microsoft-learn\": {\"type\": \"http\", \"url\": \"https://learn.microsoft.com/api/mcp\"}}}"}],"connectNote":"The endpoint answers only MCP clients over Streamable HTTP; opening it in a browser returns HTTP 405. In ChatGPT, add it as a connector after turning on developer mode in Settings, Connectors, Advanced settings. Append ?maxTokenBudget=2000 to the URL to cap search output.","faq":[{"q":"Is the Microsoft Learn MCP server free?","a":"Yes. Microsoft runs it free of charge with no API key or account (README read 2026-10-01)."},{"q":"Does the Microsoft Learn MCP server work with ChatGPT?","a":"Yes. Microsoft's README gives steps to add it in ChatGPT through developer mode connectors, and offers an OpenAI-compatible endpoint for deep research models."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/mongodb","slug":"mongodb","name":"MongoDB","title":"MongoDB MCP server: query collections and manage Atlas with AI agents","summary":"The MongoDB MCP server gives AI agents access to MongoDB databases and MongoDB Atlas: run find and aggregate queries, inspect schemas and indexes, and create or manage Atlas clusters and users.","category":"databases","publisher":"MongoDB","publisherUrl":"https://www.mongodb.com","official":true,"repo":"mongodb-js/mongodb-mcp-server","stars":1138,"repoLicense":"Apache-2.0","lastPush":"2026-10-01","registryName":"io.github.mongodb-js/mongodb-mcp-server","freeTier":"free-plan","cost":"Free and open source under Apache-2.0; database use is billed by MongoDB Atlas, which has a free cluster tier, or free on your own servers.","pricingUrl":"https://www.mongodb.com/pricing","clients":["Claude Code","Claude Desktop","Claude.ai","Codex","Cursor","VS Code","GitHub Copilot CLI","Gemini CLI","Grok Build","Devin"],"tools":["find","aggregate","count","collection-schema","collection-indexes","explain","insert-many","update-many","create-index","atlas-list-clusters","atlas-create-free-cluster","atlas-get-performance-advisor"],"whatItDoes":["The MongoDB MCP server is MongoDB's own server for connecting AI agents to MongoDB. With a connection string, the agent can list databases and collections, infer a collection's schema, run find, count and aggregation queries, explain a query plan, and read indexes and storage sizes. Write tools insert, update and delete documents, create collections and indexes, and drop them, and an export tool saves query results to a file.","MongoDB's server also manages MongoDB Atlas through the Atlas API. The agent can list and inspect organizations, projects and clusters, create a free cluster, add database users and IP access entries, load sample data, pause or upgrade a cluster, read alerts and the performance advisor, and work with Atlas Stream Processing. Atlas Local tools create and connect to a local Atlas deployment in Docker, and assistant tools search MongoDB's knowledge sources.","The server runs locally through npx with Node.js 22.13 or newer, or in Docker. MongoDB also runs a managed Atlas MCP server reached through its mongodb-atlas plugin with OAuth. All of MongoDB's setup examples start in read-only mode. The GitHub repo had 1,138 stars on 2026-10-01."],"useItWhen":["You want the agent to understand the real shape of your documents before it writes queries.","You need an aggregation pipeline written and checked against your data.","You want to create a free Atlas cluster and a database user for a new project from your editor.","You want to see slow queries and index suggestions from the Atlas performance advisor."],"notFor":["Running without read-only mode on important data: write tools can delete many documents or drop whole databases, so keep --readOnly unless you need writes."],"requirements":["Node.js 22.13 or newer, or Docker","A MongoDB connection string, or Atlas service account credentials for Atlas tools"],"connect":[{"label":"Interactive setup for the local server","command":"npx -y mongodb-mcp-server@latest setup"},{"label":"Any client with a JSON config, read-only, with a connection string","command":"{\"mcpServers\": {\"MongoDB\": {\"command\": \"npx\", \"args\": [\"-y\", \"mongodb-mcp-server@latest\", \"--readOnly\"], \"env\": {\"MDB_MCP_CONNECTION_STRING\": \"mongodb://localhost:27017/myDatabase\"}}}}"},{"label":"Codex, as a plugin from MongoDB's marketplace (then install mongodb in /plugins)","command":"codex plugin marketplace add mongodb/agent-skills"},{"label":"GitHub Copilot CLI, managed Atlas server through the plugin","command":"copilot plugin install mongodb-atlas"}],"connectNote":"For Atlas, MongoDB recommends the mongodb-atlas plugin, available in the Claude, Cursor, VS Code, Codex, Copilot CLI and Grok Build marketplaces, which signs in with OAuth. MongoDB advises passing connection strings and API credentials as environment variables, not command-line arguments.","faq":[{"q":"Is the MongoDB MCP server free?","a":"Yes. It is open source under Apache-2.0. You pay only for MongoDB itself, and Atlas has a free cluster tier (pricing page, 2026-10-01)."},{"q":"Can the MongoDB MCP server delete data?","a":"Yes, unless you run it read-only. Its tools include delete-many, drop-collection and drop-database. The --readOnly flag or MDB_MCP_READ_ONLY=true removes every create, update and delete tool."},{"q":"Does it work with self-hosted MongoDB, not only Atlas?","a":"Yes. The local server connects to any MongoDB deployment by connection string; the Atlas tools are extra and need Atlas credentials."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/neon","slug":"neon","name":"Neon","title":"Neon MCP server: serverless Postgres projects and branches for AI agents","summary":"The Neon MCP server gives AI agents control of Neon serverless Postgres: create projects and branches, run SQL, inspect schemas, find slow queries and run migrations on a branch before applying them.","category":"databases","publisher":"Neon (Databricks)","publisherUrl":"https://neon.com","official":true,"repo":"neondatabase/mcp-server-neon","stars":648,"repoLicense":"MIT","lastPush":"2026-09-30","registryName":"com.neon/mcp","remoteUrl":"https://mcp.neon.tech/mcp","freeTier":"free-plan","cost":"Free and open source under MIT; you pay only for Neon usage, and Neon has a Free plan and a pay-as-you-go plan.","pricingUrl":"https://neon.com/pricing","clients":["Claude Code","Cursor","VS Code","Kiro"],"tools":["list_projects","create_project","create_branch","describe_branch","run_sql","run_sql_transaction","get_database_tables","describe_table_schema","list_slow_queries","explain_sql_statement","prepare_database_migration","prepare_query_tuning"],"whatItDoes":["The Neon MCP server connects an AI agent to Neon, the serverless Postgres service that is now part of Databricks and built on its Lakebase Postgres. The agent can list and create projects, create, reset and delete branches, manage compute endpoints and snapshots, run SQL or a transaction, list tables and describe their schemas, and get a connection string. It also covers newer Neon products such as Neon Auth, the Data API, functions and storage.","Neon's server builds on database branching for safer changes. A migration tool first applies a schema change on a temporary branch so the agent can test it, then a second tool applies it to the main branch once you agree. A similar pair finds slow queries, tests tuning such as new indexes on a branch, and then applies the result. Other tools explain query plans and compare schemas between branches.","Neon hosts the server at mcp.neon.tech with OAuth or an API key, and URL parameters limit it to read-only tools, one project or chosen tool categories. Neon says the server is meant for local development and IDE use and does not recommend it for production. The GitHub repo had 648 stars on 2026-10-01."],"useItWhen":["You want the agent to spin up a fresh Postgres database for a prototype in one prompt.","You want schema migrations tested on a branch before they touch your main database.","A query is slow and you want the agent to find it and try an index safely.","You want a summary of all your Neon projects and what each one holds."],"notFor":["Production databases: Neon itself advises against using the MCP server there, because it can create, change and delete projects and branches.","Other Postgres hosts: it manages Neon accounts through the Neon API, not any Postgres server."],"requirements":["A Neon account (OAuth) or a Neon API key","Node.js 18 or newer for the npx setup commands"],"connect":[{"label":"One-step setup for Cursor, VS Code and Claude Code (creates an API key and configures the editor)","command":"npx neon@latest init"},{"label":"Add the hosted server with OAuth to every detected agent","command":"npx add-mcp \"https://mcp.neon.tech/mcp?category=projects&category=branches&category=endpoints&category=querying&category=schema\""},{"label":"Any client with a JSON config, read-only","command":"{\"mcpServers\": {\"Neon\": {\"url\": \"https://mcp.neon.tech/mcp?readonly=true\"}}}"}],"connectNote":"OAuth opens a browser window to authorize your Neon account; for an organization's projects, name the org_id or project_id in your prompt. With an API key, send it as Authorization: Bearer. If IP Allow is on, allow 34.192.103.46 and 23.22.233.166.","faq":[{"q":"Is the Neon MCP server free?","a":"Yes. The server is open source and costs nothing; your Neon usage is billed as normal, and Neon has a Free plan (pricing page, 2026-10-01)."},{"q":"Can the Neon MCP server delete things?","a":"Yes, with write access it can delete projects and branches. Connect with ?readonly=true or untick Allow writes during sign-in to keep it to read-only tools."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/notion","slug":"notion","name":"Notion","title":"Notion MCP server: search, read and edit Notion pages with AI agents","summary":"The Notion MCP server gives AI agents access to your Notion workspace through Notion's hosted server: search pages and databases, read them as Markdown, and create or update pages, databases and views.","category":"productivity","publisher":"Notion","publisherUrl":"https://developers.notion.com","official":true,"registryName":"com.notion/mcp","remoteUrl":"https://mcp.notion.com/mcp","freeTier":"free-plan","cost":"No separate charge, and Notion lists the MCP tools on all plans; searching connected apps through AI search needs Notion AI. The older local server is free under MIT.","pricingUrl":"https://developers.notion.com/guides/mcp/mcp-supported-tools","clients":["Claude Code","Codex","Cursor","VS Code","Google Antigravity","Devin","Hermes","Pi","fx","Claude.ai","Claude Desktop","ChatGPT"],"tools":["notion-search","notion-ai-search","notion-fetch","notion-create-pages","notion-update-page","notion-move-pages","notion-duplicate-page","notion-create-database","notion-update-data-source","notion-create-view","notion-create-file-upload","notion-get-tool-access"],"whatItDoes":["The Notion MCP server is Notion's hosted server for connecting AI agents to a Notion workspace. After you sign in with OAuth, the agent can search your workspace, fetch pages and databases in a Markdown-based format that uses few tokens, and create, update, move and duplicate pages. It can also create databases, change data source properties, set up database views with filters and sorts, and upload files up to 20 MiB.","Notion's server respects each user's existing permissions, so the agent sees only what you can see, with no integration token or manual page sharing. On workspaces with Notion AI, an AI search tool also searches connected apps alongside Notion. Workspace owners can see and revoke MCP connections in Settings, Connections, and Enterprise admins can allow or block specific MCP clients.","Notion also published an open-source local server, makenotion/notion-mcp-server, that uses an integration token. Notion now says that repo is no longer actively maintained and recommends the hosted server for most users; the local one remains an option for automated workflows that cannot complete an interactive OAuth sign-in."],"useItWhen":["You want the agent to turn meeting notes or a spec in Notion into code tasks.","You want to write documentation from your codebase straight into a Notion page.","You keep a project tracker in a Notion database and want the agent to add and update rows.","You want answers from your team wiki without copying pages into the chat."],"notFor":["Unattended automation: the hosted server needs an interactive OAuth sign-in, and Notion says non-interactive authorization is not supported yet.","Broad write access without review: the agent can edit and move any page you can edit."],"requirements":["A Notion account; Notion AI for searching connected apps"],"connect":[{"label":"Claude Code (then run /mcp and complete the OAuth flow)","command":"claude mcp add --transport http notion https://mcp.notion.com/mcp"},{"label":"Codex, in ~/.codex/config.toml (then run codex mcp login notion)","command":"[mcp_servers.notion]\nurl = \"https://mcp.notion.com/mcp\""},{"label":"Cursor, in .cursor/mcp.json","command":"{\"mcpServers\": {\"notion\": {\"url\": \"https://mcp.notion.com/mcp\"}}}"},{"label":"VS Code, in .vscode/mcp.json","command":"{\"servers\": {\"notion\": {\"type\": \"http\", \"url\": \"https://mcp.notion.com/mcp\"}}}"}],"connectNote":"Every client finishes with a Notion OAuth sign-in in the browser, where you choose the workspace. In Claude.ai, Claude Desktop and ChatGPT, Notion is added from the app's connector settings. A Notion plugin for Claude Code adds skills and slash commands on top of the server.","faq":[{"q":"Is the Notion MCP server free?","a":"There is no separate charge: Notion lists its MCP tools on all plans. Searching connected apps through AI search needs Notion AI (Notion developer docs, read 2026-10-01)."},{"q":"Does the Notion MCP server work with ChatGPT?","a":"Yes. Notion's help center names ChatGPT, Claude and Cursor as supported, and its developer docs give setup for Codex, Claude Code, Cursor, VS Code, Antigravity and Devin."},{"q":"Hosted Notion MCP or the open-source notion-mcp-server?","a":"Notion recommends the hosted server at mcp.notion.com. The open-source repo is no longer actively maintained; it is only worth using when you need token-based access without OAuth."}],"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/obsidian","slug":"obsidian","name":"Obsidian","title":"Obsidian MCP server: read, search and edit your vault with AI agents","summary":"The Obsidian MCP server gives AI agents access to your Obsidian vault through the Local REST API plugin: list and read notes, search across files, append or patch content and delete files.","category":"productivity","publisher":"Markus Pfundstein","publisherUrl":"https://github.com/MarkusPfundstein","official":false,"repo":"MarkusPfundstein/mcp-obsidian","stars":4457,"repoLicense":"MIT","lastPush":"2026-08-31","freeTier":"free","cost":"Free and open source under MIT; it needs the Obsidian app and the free Local REST API plugin, with nothing to pay.","pricingUrl":"https://github.com/MarkusPfundstein/mcp-obsidian","clients":["Claude Desktop"],"tools":["list_files_in_vault","list_files_in_dir","get_file_contents","search","patch_content","append_content","delete_file"],"whatItDoes":["mcp-obsidian is a community MCP server, written in Python by Markus Pfundstein, that lets an AI agent work inside an Obsidian vault. It does not open the vault files directly: it calls the Local REST API community plugin, which runs inside the Obsidian app on your machine and listens on port 27124 by default. The agent can list folders, read a note, run a text search across the whole vault, and write back.","The mcp-obsidian write tools cover three cases. append_content adds text to a new or existing note, patch_content inserts text next to a heading, a block reference or a frontmatter field, and delete_file removes a file or folder. That is enough for jobs like turning meeting notes into a summary note or adding a task under a heading, without the agent rewriting whole files.","mcp-obsidian runs locally with uvx or in Docker, and it needs Obsidian to be open with the plugin enabled, because the plugin is the only way in. The README pins the mcp Python SDK below 2.0, since the server uses the 1.x low-level API that 2.0 removed. The repo had 4,456 GitHub stars on 2026-10-01; its last push was on 2026-08-31."],"useItWhen":["You want the agent to find every note that mentions a topic and explain the context.","You want a summary of last week's meeting notes written into a new note.","You keep project docs in Obsidian and want the agent to read them while it codes.","You want to add tasks or links under a specific heading without opening the app."],"notFor":["Vaults you cannot afford to lose: the server can delete files and folders, so keep a backup or Obsidian Sync history before letting an agent write.","Use without the Obsidian app running: every call goes through the Local REST API plugin, so the app must be open on the same machine or reachable on the network.","Official support: the server is community-maintained and not published by Obsidian."],"requirements":["Obsidian with the Local REST API community plugin installed and enabled, and its API key","Python 3.11 or newer with uv (uvx), or Docker"],"connect":[{"label":"Any client with a JSON config (Claude Desktop shown in the README)","command":"{\"mcpServers\": {\"mcp-obsidian\": {\"command\": \"uvx\", \"args\": [\"mcp-obsidian\"], \"env\": {\"OBSIDIAN_API_KEY\": \"<YOUR_OBSIDIAN_API_KEY>\", \"OBSIDIAN_HOST\": \"<your_obsidian_host>\", \"OBSIDIAN_PORT\": \"<your_obsidian_port>\"}}}}"},{"label":"Docker, after copying .env.example to .env and building the image","command":"docker compose run --rm -T mcp-obsidian"}],"connectNote":"Copy the API key from the Local REST API plugin settings in Obsidian. Host defaults to 127.0.0.1 and port to 27124 when not set. The README only documents Claude Desktop, but the same JSON block works in any client that runs local stdio servers; if a client cannot find uvx, put the full path from `which uvx` in the command.","faq":[{"q":"Is there an official Obsidian MCP server?","a":"No. Obsidian does not publish one as of 2026-10-01. mcp-obsidian by Markus Pfundstein is a community server with 4,456 GitHub stars, and the MCP Registry lists other community Obsidian servers."},{"q":"Does the Obsidian MCP server work without the Obsidian app open?","a":"No. It sends every request to the Local REST API plugin, which runs inside Obsidian, so the app must be running with the plugin enabled."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/perplexity","slug":"perplexity","name":"Perplexity","title":"Perplexity MCP server: web search, answers and deep research for AI agents","summary":"The Perplexity MCP server gives AI agents Perplexity's web search, quick answers with citations, step-by-step reasoning and long research runs, all billed to a Perplexity API account.","category":"web-search-and-scraping","publisher":"Perplexity","publisherUrl":"https://www.perplexity.ai","official":true,"repo":"perplexityai/modelcontextprotocol","stars":2550,"repoLicense":"MIT","lastPush":"2026-09-25","registryName":"ai.perplexity/mcp-server","remoteUrl":"https://api.perplexity.ai/mcp","freeTier":"paid","cost":"The server is free and open source under MIT, but every call is billed to your Perplexity API account, which has no free tier.","pricingUrl":"https://docs.perplexity.ai/docs/getting-started/pricing","clients":["Claude Code","Codex","Cursor","VS Code","Claude Desktop","Claude.ai","Google Antigravity","Kiro","Windsurf"],"tools":["perplexity_search","perplexity_ask","perplexity_reason","perplexity_research"],"whatItDoes":["The Perplexity MCP server connects an AI agent to the Perplexity API Platform. It adds four tools that differ in depth. The search tool returns ranked web results with metadata from the Search API, with filters for recency and domains. The ask tool answers a question in a conversational way using live web search, which suits quick lookups.","The other two tools take longer. The reason tool works through an analytical problem step by step, and the research tool runs a deep, multi-step investigation that can last minutes and returns a detailed report; the server streams progress to clients that ask for it. Current versions run these tools on Perplexity's Agent API presets, which Perplexity tunes over time, instead of the older Sonar model names.","Perplexity hosts the server at api.perplexity.ai/mcp, reached by OAuth sign-in or an API key, and publishes the same code on npm to run locally or in Docker. The GitHub repo had 2,549 stars on 2026-10-01. It is also packaged as a Claude Code plugin and an Agent Plugin."],"useItWhen":["You want an agent to get a cited answer from the web in one call, not a list of links.","You need a long research report on a topic and can wait a few minutes for it.","You want plain ranked search results filtered to recent pages or chosen domains.","Your team already pays for the Perplexity API and wants it inside its editor."],"notFor":["Free use: there is no free allowance on the API pricing page, so every search and answer is billed, and research runs can cost more than a simple search."],"requirements":["A Perplexity API account (OAuth sign-in) or API key from the Perplexity console","Node.js, only when running the server locally"],"connect":[{"label":"Claude Code, remote server (sign in through /mcp)","command":"claude mcp add --transport http perplexity https://api.perplexity.ai/mcp"},{"label":"Claude Code, remote server with an API key","command":"claude mcp add --transport http perplexity https://api.perplexity.ai/mcp --header \"Authorization: Bearer YOUR_API_KEY\""},{"label":"Codex, local server","command":"codex mcp add perplexity --env PERPLEXITY_API_KEY=\"your_key_here\" -- npx -y @perplexity-ai/mcp-server"},{"label":"Cursor, Claude Desktop, Windsurf, Kiro or VS Code, local server","command":"{\"mcpServers\": {\"perplexity\": {\"command\": \"npx\", \"args\": [\"-y\", \"@perplexity-ai/mcp-server\"], \"env\": {\"PERPLEXITY_API_KEY\": \"your_key_here\"}}}}"}],"connectNote":"With OAuth, the first call opens a browser where you sign in to Perplexity and choose which API organization pays. In Claude.ai, add https://api.perplexity.ai/mcp as a custom connector. Strict clients that fail on startup can use npx -yq instead of npx -y.","faq":[{"q":"Is the Perplexity MCP server free?","a":"The server code is free, but calls are not: they are billed to the Perplexity API, per request for search and per token for answers and research (pricing docs, 2026-10-01). A Perplexity Pro chat subscription is a separate product."},{"q":"Which Perplexity MCP tool should an agent use?","a":"perplexity_search for raw results, perplexity_ask for a quick cited answer, perplexity_reason for analysis, and perplexity_research for a long report that can take minutes."},{"q":"Does it work in Codex and Cursor?","a":"Yes. Perplexity's README gives a codex mcp add command and a Cursor one-click install, plus config for Claude Desktop, Windsurf, Kiro and VS Code."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/playwright","slug":"playwright","name":"Playwright","title":"Playwright MCP server: browser automation for AI agents","summary":"The Playwright MCP server gives AI agents a real browser driven by Microsoft's Playwright: navigate, click, type and fill forms using the page's accessibility tree instead of screenshots.","category":"browser-automation","publisher":"Microsoft","publisherUrl":"https://playwright.dev","official":true,"repo":"microsoft/playwright-mcp","stars":37745,"repoLicense":"Apache-2.0","lastPush":"2026-09-28","registryName":"io.github.microsoft/playwright-mcp","freeTier":"free","cost":"Free and open source under Apache-2.0; it runs a browser on your own machine, with no account, key or usage bill.","pricingUrl":"https://github.com/microsoft/playwright-mcp","clients":["Claude Code","Claude Desktop","Codex","GitHub Copilot CLI","VS Code","Cursor","Windsurf","Cline","Gemini CLI","Goose","Grok","Junie","Kiro","LM Studio","opencode","Amp","Factory","Qodo Gen","Warp","Google Antigravity"],"tools":["browser_navigate","browser_snapshot","browser_click","browser_type","browser_fill_form","browser_select_option","browser_press_key","browser_wait_for","browser_take_screenshot","browser_evaluate","browser_console_messages","browser_tabs"],"whatItDoes":["The Playwright MCP server is Microsoft's browser automation server for AI agents, built on the Playwright testing library. It opens Chrome, Edge, Firefox or WebKit and lets the agent navigate, click, type, hover, drag, upload files, fill forms and handle dialogs. The agent reads each page as a structured accessibility snapshot rather than an image, so it does not need a vision model and its clicks hit exact elements.","Beyond the core tools, Playwright groups extra abilities behind opt-in flags: network mocking and request inspection, cookie and local storage control, PDF export, coordinate-based mouse control for canvas apps, test assertions that check text or elements are visible, and DevTools features such as tracing and video recording. The browser keeps a persistent profile per project by default, can run isolated for clean test sessions, or attach to your own logged-in Chrome or Edge through the Playwright extension.","The server runs locally through npx with Node.js 18 or newer, and also ships as a Docker image. The GitHub repo had 37,737 stars on 2026-10-01. Microsoft's README notes that coding agents may spend fewer tokens with its separate Playwright CLI plus skills, and keeps MCP for long, stateful browser loops."],"useItWhen":["You want the agent to click through your web app and check that a feature works.","You need to fill in a multi-step web form or pull data from pages that need a login.","You want the agent to write or repair Playwright end-to-end tests by trying the flow first.","You need to mock an API response or block requests while the agent tests a page."],"notFor":["Token-tight coding sessions: accessibility snapshots of large pages are long, and Microsoft itself points coding agents to the lighter Playwright CLI.","Debugging performance or network timing in detail; Chrome DevTools MCP has deeper profiling."],"requirements":["Node.js 18 or newer","A local browser (Playwright can install one)"],"connect":[{"label":"Claude Code","command":"claude mcp add playwright npx @playwright/mcp@latest"},{"label":"Codex","command":"codex mcp add playwright npx \"@playwright/mcp@latest\""},{"label":"VS Code","command":"code --add-mcp '{\"name\":\"playwright\",\"command\":\"npx\",\"args\":[\"@playwright/mcp@latest\"]}'"},{"label":"Standard config for Cursor, Claude Desktop, Windsurf and most clients","command":"{\"mcpServers\": {\"playwright\": {\"command\": \"npx\", \"args\": [\"@playwright/mcp@latest\"]}}}"}],"connectNote":"Add --headless to run without a visible window, --isolated for a fresh profile each session, or --extension to drive your existing Chrome or Edge tabs through the Playwright extension.","faq":[{"q":"Is the Playwright MCP server free?","a":"Yes. It is open source under Apache-2.0 and runs on your machine, with no account or usage fees."},{"q":"Does Playwright MCP need screenshots or a vision model?","a":"No. It gives the agent the page's accessibility tree as text. A vision mode with coordinate clicks exists as an opt-in for canvas-style pages."},{"q":"Can Playwright MCP use my logged-in browser?","a":"Yes, in two ways: its default persistent profile keeps logins between sessions, and the --extension flag connects to tabs in your own Chrome or Edge."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/postgres-mcp-pro","slug":"postgres-mcp-pro","name":"Postgres MCP Pro","title":"Postgres MCP Pro: PostgreSQL queries and index tuning for AI agents","summary":"Postgres MCP Pro is an open-source MCP server that gives AI agents access to a PostgreSQL database, with schema browsing, safe SQL execution, EXPLAIN plans, index recommendations and database health checks.","category":"databases","publisher":"Crystal DBA","publisherUrl":"https://www.crystaldba.ai","official":false,"repo":"crystaldba/postgres-mcp","stars":3361,"repoLicense":"MIT","lastPush":"2026-08-17","freeTier":"free","cost":"Free and open source under MIT, with no paid tier; it runs on your machine against your own PostgreSQL database.","pricingUrl":"https://github.com/crystaldba/postgres-mcp","clients":["Claude Desktop","Cursor","Windsurf","Cline","Goose","Qodo Gen"],"tools":["list_schemas","list_objects","get_object_details","execute_sql","explain_query","get_top_queries","analyze_workload_indexes","analyze_query_indexes","analyze_db_health"],"whatItDoes":["Postgres MCP Pro is an MCP server from Crystal DBA, a company that builds tools for PostgreSQL. It connects an AI agent to any Postgres database by connection URI. The agent can list schemas and objects, read a table's columns, constraints and indexes, run SQL, and get the execution plan for a query, which gives it enough context to write SQL that fits the real schema.","Postgres MCP Pro's main difference from a plain database connector is tuning. It reports the slowest queries from pg_stat_statements, recommends indexes for the whole workload or for up to 10 given queries by searching many index combinations, and uses the hypopg extension to test hypothetical indexes in EXPLAIN without creating them. A health check covers buffer cache hit rates, connections, vacuum, sequence limits, replication lag and duplicate or unused indexes.","The server runs locally through Docker, uvx, pipx or uv and needs Python 3.12 or newer outside Docker. It has two access modes: unrestricted for development, and restricted, which allows only read-only transactions with an execution time limit, for production. The repo had 3,359 GitHub stars on 2026-10-01; its last push was 2026-08-17."],"useItWhen":["Your app's queries are slow and you want index suggestions backed by the real workload.","You want the agent to read your schema before it writes SQL or ORM code.","You want a health report on a production database with no write access for the agent.","You want to check whether a planned index would change a query plan before adding it."],"notFor":["Databases other than PostgreSQL; DBHub covers MySQL, SQL Server, SQLite and others.","Unattended use in unrestricted mode: the agent can then change data and schema, so use --access-mode=restricted on anything that matters."],"requirements":["A PostgreSQL connection URI","Docker, or Python 3.12 or newer","The pg_stat_statements and hypopg extensions for index tuning and workload analysis"],"connect":[{"label":"Claude Desktop or any client with a JSON config, using Docker","command":"{\"mcpServers\": {\"postgres\": {\"command\": \"docker\", \"args\": [\"run\", \"-i\", \"--rm\", \"-e\", \"DATABASE_URI\", \"crystaldba/postgres-mcp\", \"--access-mode=unrestricted\"], \"env\": {\"DATABASE_URI\": \"postgresql://localhost:5432/dbname\"}}}}"},{"label":"The same with uvx, in read-only restricted mode","command":"{\"mcpServers\": {\"postgres\": {\"command\": \"uvx\", \"args\": [\"postgres-mcp\", \"--access-mode=restricted\"], \"env\": {\"DATABASE_URI\": \"postgresql://localhost:5432/dbname\"}}}}"},{"label":"Shared server over SSE with Docker","command":"docker run -p 8000:8000 -e DATABASE_URI=postgresql://localhost:5432/dbname crystaldba/postgres-mcp --access-mode=unrestricted --transport=sse"}],"connectNote":"Put your database user and password into the URI in the usual user:password@host form. The Docker image remaps localhost to the host machine automatically. On AWS RDS, Azure or Google Cloud SQL, enable pg_stat_statements and hypopg through the provider before using the tuning tools.","faq":[{"q":"Is Postgres MCP Pro free?","a":"Yes. It is open source under MIT with no paid version; you run it yourself against your own database."},{"q":"Can Postgres MCP Pro write to my database, or only read?","a":"That depends on the mode. Unrestricted mode allows full reads and writes, including schema changes; restricted mode limits the agent to read-only transactions with a time limit."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/redis","slug":"redis","name":"Redis","title":"Redis MCP server: read, write and vector-search Redis data with AI agents","summary":"The Redis MCP server gives AI agents natural-language access to a Redis database: strings, hashes, lists, sets, sorted sets, streams, JSON documents, pub/sub and vector search indexes.","category":"databases","publisher":"Redis","publisherUrl":"https://redis.io","official":true,"repo":"redis/mcp-redis","stars":629,"repoLicense":"MIT","lastPush":"2026-09-21","registryName":"io.github.redis/mcp-redis","freeTier":"free-plan","cost":"Free and open source under MIT and works with any Redis you run; Redis Cloud, the hosted option, has a free database and paid plans.","pricingUrl":"https://redis.io/pricing/","clients":["Claude Desktop","VS Code with GitHub Copilot","Augment","OpenAI Agents SDK"],"whatItDoes":["The Redis MCP server is Redis's own MCP server for working with data in a Redis database. It gives an AI agent tools for every main Redis data type: set and get strings with expiry, store objects as hashes, push and pop list items for queues, manage sets and sorted sets for tags or leaderboards, and add to and read streams with consumer groups. JSON tools store and edit nested documents by path.","Redis's server also covers newer uses of Redis in AI apps. Query engine tools create vector indexes and run vector similarity search, and hashes can hold embeddings. Pub/sub tools publish messages and read from subscriptions. A docs tool answers questions from Redis documentation, and a server tool reports database information.","The server is a Python package that runs locally with uvx or Docker over stdio; Redis says Streamable HTTP support will come later. It connects by redis:// or rediss:// URL, supports Redis ACL users, and can sign in to Azure Managed Redis with Microsoft Entra ID. The GitHub repo had 628 stars on 2026-10-01."],"useItWhen":["You want to look inside a Redis cache or queue while debugging, in plain words.","You are building an agent that needs short-term memory or session storage in Redis.","You want to create a vector index and test similarity search without writing client code.","You use Azure Managed Redis and want Entra ID sign-in instead of passwords."],"notFor":["Production caches without a limited ACL user: the agent can overwrite and delete keys, so give it a Redis user with only the commands it needs.","Clients that only speak HTTP: the server supports stdio only for now."],"requirements":["Python with uv (uvx), or Docker","A Redis server or Redis Cloud database URL"],"connect":[{"label":"Claude Desktop or any client with a JSON config (use the full path to uvx in Claude Desktop)","command":"{\"mcpServers\": {\"redis-mcp-server\": {\"type\": \"stdio\", \"command\": \"uvx\", \"args\": [\"--from\", \"redis-mcp-server@latest\", \"redis-mcp-server\", \"--url\", \"redis://localhost:6379/0\"]}}}"},{"label":"VS Code with GitHub Copilot, in mcp.json","command":"\"servers\": {\"redis\": {\"type\": \"stdio\", \"command\": \"uvx\", \"args\": [\"--from\", \"redis-mcp-server@latest\", \"redis-mcp-server\", \"--url\", \"redis://localhost:6379/0\"]}}"}],"connectNote":"For Redis Cloud or any TLS connection, use a rediss:// URL with your user and password. VS Code needs agent mode turned on (chat.agent.enabled). Redis recommends a Redis ACL user with limited rights for the agent.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/sentry","slug":"sentry","name":"Sentry","title":"Sentry MCP server: errors, traces and issue triage for AI agents","summary":"The Sentry MCP server gives AI coding agents access to your Sentry account: search errors and events, read issue details and stack traces, look at performance data and triage issues while fixing code.","category":"cloud-and-devops","publisher":"Sentry","publisherUrl":"https://sentry.io","official":true,"repo":"getsentry/toolkit","stars":882,"repoLicense":"FSL-1.1-Apache-2.0","lastPush":"2026-10-01","registryName":"io.github.getsentry/sentry-mcp","remoteUrl":"https://mcp.sentry.dev/mcp","freeTier":"free-plan","cost":"No extra charge: Sentry lists MCP access on its free Developer plan; paid plans add volume and users, and Seer, Sentry's AI debugger, costs extra.","pricingUrl":"https://sentry.io/pricing/","clients":["Claude Code","Cursor","VS Code"],"tools":["search_issues","search_events","get_issue_details","get_event_stacktrace","get_trace_details","update_issue","add_issue_note","analyze_issue_with_seer","find_projects","find_releases","search_docs","create_project"],"whatItDoes":["The Sentry MCP server connects an AI coding agent to Sentry, the error and performance monitoring service. Sentry built it for developers who debug with an agent at their side: the agent can look up the issues in a project, read an error's stack trace and events, search events and traces, check performance problems and read Sentry's own documentation. Project management tools let it assign or resolve issues and manage projects, and a Seer group hands an issue to Sentry's AI root-cause analysis.","Sentry's search tools accept plain-language questions, such as \"errors in checkout in the last day\", and translate them into Sentry's query syntax with a language model. On the hosted server Sentry handles that; on a self-run server you must configure an OpenAI, Azure OpenAI, Anthropic or OpenRouter key for those tools to work. Tools are grouped into skills that you can narrow with a URL parameter, and scoping the URL to one organization or project hides discovery tools the agent does not need.","Sentry hosts the server at mcp.sentry.dev with OAuth sign-in. The code also runs locally with npx, which is the way to connect a self-hosted Sentry. The source repo was renamed from sentry-mcp to getsentry/toolkit and had 875 GitHub stars on 2026-10-01; it uses the Functional Source License, which converts to Apache-2.0 after two years."],"useItWhen":["You want the agent to read a production error's stack trace and fix the code that caused it.","You want a list of the newest or most frequent issues in a project before a release.","You want to resolve or assign Sentry issues without leaving your editor.","You run self-hosted Sentry and want the same tools against your own instance."],"notFor":["General Sentry administration: Sentry says it focuses the tools on debugging workflows, not on covering the whole Sentry API.","Self-run setups without an LLM key: the plain-language search tools need an OpenAI, Anthropic or OpenRouter key when you run the server yourself."],"requirements":["A Sentry account (OAuth) or a Sentry user auth token for the local server"],"connect":[{"label":"Claude Code, scoped to one project","command":"claude mcp add --transport http sentry https://mcp.sentry.dev/mcp/{organizationSlug}/{projectSlug}"},{"label":"Claude Code, as a plugin with a Sentry subagent","command":"claude plugin marketplace add getsentry/sentry-mcp && claude plugin install sentry-mcp@sentry-mcp"},{"label":"Cursor and other clients with a JSON config","command":"{\"mcpServers\": {\"sentry\": {\"url\": \"https://mcp.sentry.dev/mcp/{organizationSlug}/{projectSlug}\"}}}"},{"label":"Local server, for example against self-hosted Sentry","command":"npx @sentry/mcp-server@latest --access-token=sentry-user-token --host=sentry.example.com"}],"connectNote":"The first connection to the hosted server opens a Sentry OAuth sign-in. In VS Code, run MCP: Add Server, choose HTTP and paste the endpoint. Add ?experimental=1 to the URL for newer tool variants.","faq":[{"q":"Is the Sentry MCP server free?","a":"Using it costs nothing extra: Sentry lists MCP access on its free Developer plan. You pay only for your Sentry plan and for Seer if you use its AI analysis (pricing page, 2026-10-01)."},{"q":"Can the Sentry MCP server change anything, or only read?","a":"It can change things: it can update issues, for example to assign or resolve them, and manage projects. Use the skills parameter to expose only the read and inspect tools."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/slack","slug":"slack","name":"Slack","title":"Slack MCP server: search, read and send Slack messages for AI agents","summary":"The Slack MCP server gives AI agents access to a Slack workspace through Slack's hosted endpoint: search messages and files, read channels and threads, send messages, and edit canvases and lists.","category":"productivity","publisher":"Slack","publisherUrl":"https://slack.com","official":true,"remoteUrl":"https://mcp.slack.com/mcp","freeTier":"free-plan","cost":"Slack's MCP docs list no separate charge and name no plan limits; your Slack plan is billed as usual, and workspace admins approve access.","pricingUrl":"https://docs.slack.dev/ai/slack-mcp-server/","clients":["Claude.ai","Claude Code","Cursor","Perplexity"],"whatItDoes":["The Slack MCP server is Slack's own hosted server for connecting AI agents to a workspace, generally available since 2026-02-17. Through it an agent can search messages, files, users, channels and emoji, read a channel or a thread, send or schedule a message, add reactions, create and update canvases and lists, upload files and read user profiles. Each action runs as the user who signed in.","The Slack MCP server only accepts connections through Slack apps that are either published in the Slack Marketplace or built internally by your organization, using user-token OAuth with a client ID and secret. In practice that means you connect through a partner client whose app your workspace admins approve, or build an internal app. Admins can review and manage every MCP client integration.","Slack names Claude.ai, Claude Code, Cursor and Perplexity as supported clients in its docs, and its launch post lists OpenAI, Google and Anthropic among the partners. Rate limits follow Slack's API tiers, with special limits on search and sending. Slack publishes no source code, so there is no GitHub repo or self-hosted version."],"useItWhen":["You want a summary of a long thread or a busy channel you missed.","You want the agent to find the message where a decision was made.","You want a status update posted to a channel when the agent finishes a task.","You want meeting notes turned into a Slack canvas."],"notFor":["A quick personal connection without admin approval: the server only works through Marketplace or internal Slack apps that workspace admins approve.","Unwatched agents: the server can post messages and edit canvases as you, so keep human confirmation on for send tools.","Self-hosting: if you need a local server, korotovsky/slack-mcp-server is the most-starred community option (1,852 GitHub stars on 2026-10-02, MIT, last push 2026-07-16), but it is not made by Slack."],"requirements":["A Slack workspace where admins allow the MCP client's app","A supported MCP client, or an internal Slack app with user-token OAuth"],"connect":[{"label":"Claude Code, Slack's plugin with the MCP server and skills (type in a Claude Code session)","command":"/plugin install slack@claude-plugins-official"},{"label":"Cursor, Slack's plugin with the MCP server and skills (type in Cursor's agent chat)","command":"/add-plugin slack"},{"label":"Server URL for other supported clients (Claude.ai, Perplexity)","command":"https://mcp.slack.com/mcp"}],"connectNote":"Both plugins set up the server and ask for Slack OAuth on first use; a workspace admin must approve the MCP integration first. The plugin's Codex install (codex plugin add slack@slack) adds the skills only, because Slack has not wired the MCP server into Codex yet. For your own agent, create an internal Slack app, use its client ID and secret in the OAuth flow, and request only the scopes your tools need, such as search:read.public or chat:write.","faq":[{"q":"Is there an official Slack MCP server?","a":"Yes. Slack made its hosted MCP server at mcp.slack.com/mcp generally available on 2026-02-17. It works through Marketplace or internal Slack apps with OAuth, not with a pasted token."},{"q":"Can the Slack MCP server send messages?","a":"Yes. It can send and schedule messages, add reactions, upload files and edit canvases and lists, in addition to search and reading channels and threads. The chat:write scope controls sending."}],"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/stripe","slug":"stripe","name":"Stripe","title":"Stripe MCP server: payments, billing and Stripe API for AI agents","summary":"The Stripe MCP server gives AI agents access to a Stripe account through Stripe's hosted endpoint: read and write most of the Stripe API, search the docs and plan a Stripe integration.","category":"payments","publisher":"Stripe","publisherUrl":"https://stripe.com","official":true,"repo":"stripe/ai","stars":1852,"repoLicense":"MIT","lastPush":"2026-10-02","registryName":"com.stripe/mcp","remoteUrl":"https://mcp.stripe.com","freeTier":"free-plan","cost":"No separate charge for the server; payments, invoices and refunds the agent creates are billed at your normal Stripe fees, and analytics queries need Stripe Sigma.","pricingUrl":"https://docs.stripe.com/mcp","clients":["Claude Code","Claude Desktop","Claude.ai","Codex","ChatGPT","Cursor","VS Code","Grok Build"],"tools":["stripe_api_search","stripe_api_details","stripe_api_read","stripe_api_write","get_stripe_account_info","stripe_analytics","get_balance_summary","search_stripe_documentation","stripe_implementation_planner","send_stripe_feedback"],"whatItDoes":["The Stripe MCP server is Stripe's own hosted server for connecting AI agents to a Stripe account. Instead of one tool per endpoint, it gives the agent four API tools: search for an API method, read its parameters, then call it with stripe_api_read for GET requests or stripe_api_write for POST, PATCH, PUT and DELETE. That covers customers, payments, refunds, invoices, subscriptions, products, prices, coupons, payment links, disputes, tax and more.","The Stripe server also searches Stripe's docs and support articles, plans an integration step by step, and in preview queries metrics and runs financial reports. Stripe asks a human to confirm certain writes, such as refunds and outbound payments, through a link the agent shows you; unconfirmed requests expire after 24 hours. Connect platforms can act for a connected account with a restricted key and the Stripe-Account header.","Stripe hosts the server at mcp.stripe.com with OAuth, where you pick live or sandbox accounts and permissions per environment, or with an agent API key. Admins can turn MCP access on or off for the whole team. The stripe/ai repo, formerly agent-toolkit, holds Stripe's AI SDKs, plugins and skills, and had 1,851 GitHub stars on 2026-10-01."],"useItWhen":["You want the agent to look up a customer's payments and subscription status.","You want test products, prices and payment links created in a sandbox.","You are building a Stripe integration and want the agent to read the current docs.","You want a refund issued with a human approval step."],"notFor":["Live mode without limits: stripe_api_write can create charges, cancel subscriptions and issue refunds, so start in a sandbox and grant live access per permission.","Old API keys after 2026-10-31: from that date Stripe MCP rejects full secret keys and restricted keys without the Agent tag, so switch to an agent key or OAuth."],"requirements":["A Stripe account, signed in through OAuth or with an agent API key","Stripe Sigma, only for natural language SQL queries in stripe_analytics"],"connect":[{"label":"Stripe plugin and skills for detected agents, step 1 (recommended by Stripe)","command":"npm install -g @stripe/cli@latest"},{"label":"Stripe plugin and skills for detected agents, step 2","command":"stripe agent setup"},{"label":"Claude Code (then run claude /mcp to sign in)","command":"claude mcp add --transport http stripe https://mcp.stripe.com/"},{"label":"Codex","command":"codex mcp add stripe --url https://mcp.stripe.com"},{"label":"VS Code .vscode/mcp.json","command":"{\"servers\": {\"stripe\": {\"type\": \"http\", \"url\": \"https://mcp.stripe.com\"}}}"}],"connectNote":"Use OAuth when your client supports it. For an autonomous agent, create an agent API key with only the permissions it needs and pass it as a bearer token in the Authorization header, read from an environment variable rather than pasted into a command. Claude Desktop and ChatGPT install Stripe from their connector and plugin directories.","faq":[{"q":"Is the Stripe MCP server free?","a":"Stripe lists no charge for the MCP server itself (docs read 2026-10-01). Anything the agent does in your account, such as creating payments or subscriptions, is billed at your normal Stripe fees."},{"q":"Can the Stripe MCP server move money?","a":"Yes, within the permissions you grant. stripe_api_write can create charges, refunds and subscriptions. Stripe requires human confirmation for some writes, such as refunds and outbound payments, and sandbox access lets you test first."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/supabase","slug":"supabase","name":"Supabase","title":"Supabase MCP server: manage Postgres and migrations with AI agents","summary":"The Supabase MCP server gives AI agents access to your Supabase projects: list tables, run SQL, apply migrations, read logs and advisors, deploy edge functions, manage branches and search Supabase docs.","category":"databases","publisher":"Supabase","publisherUrl":"https://supabase.com","official":true,"repo":"supabase/mcp","stars":2927,"repoLicense":"Apache-2.0","lastPush":"2026-10-01","registryName":"com.supabase/mcp","remoteUrl":"https://mcp.supabase.com/mcp","freeTier":"free-plan","cost":"The server costs nothing extra and is open source under Apache-2.0; it works with any Supabase plan, including the Free plan.","pricingUrl":"https://supabase.com/pricing","clients":["Claude Code","Claude.ai","ChatGPT","Codex","Cursor","VS Code","GitHub Copilot","Gemini CLI","Grok","OpenCode","Goose","Google Antigravity","Kiro","Devin Desktop","Kimi Code","Warp","Factory","fx"],"tools":["list_projects","list_tables","execute_sql","apply_migration","list_migrations","list_extensions","get_advisors","generate_typescript_types","deploy_edge_function","create_branch","merge_branch","search_docs"],"whatItDoes":["The Supabase MCP server connects an AI agent to Supabase, the Postgres-based backend platform. Through it an agent can list your projects and tables, run SQL, write and apply migrations, list extensions, generate TypeScript types from the schema, and read the security and performance advisors. It can also read logs, deploy edge functions, fetch project URLs and keys, and search Supabase's documentation.","Supabase's server ties into its development branches: an agent can create a branch database, test a migration there and merge it back, which keeps experiments off production. URL parameters narrow what the agent can reach: project_ref limits it to one project, read_only=true runs SQL as a read-only Postgres user, and features picks tool groups such as database or docs. Supabase also wraps SQL results with instructions meant to discourage the model from following commands hidden in your data.","Supabase hosts the server at mcp.supabase.com with OAuth sign-in, and a local Supabase CLI stack serves a smaller tool set at localhost:54321/mcp. The open-source package also lets you serve the tools from your own endpoint. The GitHub repo had 2,925 stars on 2026-10-01."],"useItWhen":["You build on Supabase and want the agent to design tables and write migrations against the real schema.","You want the agent to read the security advisor and fix missing row level security policies.","You need to deploy or debug an edge function from your editor.","You want schema experiments to happen on a branch database, not on production."],"notFor":["Giving to customers or end users: Supabase warns the server runs with your developer permissions.","Production data with full access: prompt injection in stored data could trick the agent into harmful queries, so use read_only=true and project_ref there."],"requirements":["A Supabase account (OAuth sign-in on first use)"],"connect":[{"label":"Claude Code, project config (then run claude /mcp and choose Authenticate)","command":"claude mcp add --scope project --transport http supabase \"https://mcp.supabase.com/mcp\""},{"label":"Codex (then run codex mcp login supabase)","command":"codex mcp add supabase --url \"https://mcp.supabase.com/mcp\""},{"label":"Gemini CLI","command":"gemini mcp add -t http supabase \"https://mcp.supabase.com/mcp\""},{"label":"Cursor, VS Code and most clients","command":"{\"mcpServers\": {\"supabase\": {\"type\": \"http\", \"url\": \"https://mcp.supabase.com/mcp\"}}}"}],"connectNote":"Append ?project_ref=<id>&read_only=true to the URL to limit the agent to one project with read-only SQL. The connection tab in the Supabase dashboard builds a custom URL for your project.","faq":[{"q":"Is the Supabase MCP server free?","a":"Yes. The server has no separate price and works on Supabase's Free plan; you pay only if you move to a paid Supabase plan (pricing page, 2026-10-01)."},{"q":"Can the Supabase MCP server write to my database?","a":"Yes by default: it can run any SQL, apply migrations and deploy functions. Add read_only=true to the URL to run queries as a read-only Postgres user and drop the mutating tools."},{"q":"Does the Supabase MCP server work with ChatGPT?","a":"Yes. Supabase's setup docs list ChatGPT and Claude.ai among web clients, along with Codex, Claude Code, Cursor, VS Code, Gemini CLI and Grok."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/tavily","slug":"tavily","name":"Tavily","title":"Tavily MCP server: real-time web search and extraction for AI agents","summary":"The Tavily MCP server gives AI agents real-time web search through the Tavily API, plus tools that extract page content, map a site's structure and crawl it.","category":"web-search-and-scraping","publisher":"Tavily","publisherUrl":"https://www.tavily.com","official":true,"repo":"tavily-ai/tavily-mcp","stars":2415,"repoLicense":"MIT","lastPush":"2026-09-29","registryName":"io.github.tavily-ai/tavily-mcp","remoteUrl":"https://mcp.tavily.com/mcp","freeTier":"free-plan","cost":"Free and open source under MIT; calls spend Tavily API credits, with a free monthly allowance, then pay as you go or a paid plan.","pricingUrl":"https://www.tavily.com/pricing","clients":["Claude Code","Claude Desktop","Cursor","Cline"],"tools":["tavily_search","tavily_extract","tavily_map","tavily_crawl","tavily_research"],"whatItDoes":["The Tavily MCP server connects an AI agent to Tavily, a search API made for language models. Its search tool queries the live web and returns ranked results with content snippets, with options for search depth, number of results, images and time range. Its extract tool pulls the readable content out of one or more URLs, so the agent can read a source after finding it.","Tavily's server also covers whole sites. A map tool returns the structure of a website as a list of its pages, and a crawl tool walks a site from a starting URL and collects content along the way. The source code also registers a research tool for longer multi-step questions. Default parameters such as search depth or result count can be set once for every call, through an environment variable locally or a header on the remote server.","Tavily hosts the server at mcp.tavily.com and supports OAuth sign-in or an API key, so most people connect a URL. The same server runs locally with npx and Node.js 20 or newer. The GitHub repo had 2,415 stars on 2026-10-01."],"useItWhen":["Your agent needs fresh search results with the page text, not only links.","You want to pull the content of a list of URLs into the conversation.","You need a list of every page on a documentation site before reading some of them.","You want a search provider with a free monthly allowance for a side project."],"notFor":["Use without an account: every call needs a Tavily API key or OAuth sign-in, and the free plan has a monthly credit limit."],"requirements":["A Tavily account or API key","Node.js 20 or newer, only when running the server locally"],"connect":[{"label":"Claude Code, with OAuth sign-in (finish it with /mcp)","command":"claude mcp add --transport http tavily https://mcp.tavily.com/mcp"},{"label":"Claude Code, with an API key in the URL","command":"claude mcp add --transport http tavily https://mcp.tavily.com/mcp/?tavilyApiKey=<your-api-key>"},{"label":"Cursor, in mcp.json","command":"{\"mcpServers\": {\"tavily-remote-mcp\": {\"command\": \"npx -y mcp-remote https://mcp.tavily.com/mcp/?tavilyApiKey=<your-api-key>\", \"env\": {}}}}"},{"label":"Run the server locally (set TAVILY_API_KEY in its environment)","command":"npx -y tavily-mcp@latest"}],"connectNote":"Tavily accepts the API key in the URL, as an Authorization: Bearer header, or through OAuth. In Claude Desktop, add the remote URL as a custom connector.","faq":[{"q":"Is the Tavily MCP server free?","a":"The server is free and open source, and Tavily's free Researcher plan includes a monthly allowance of API credits. After that, you pay as you go or move to a monthly plan (pricing page, 2026-10-01)."},{"q":"Do I need an API key for the Tavily MCP server?","a":"Yes, or an OAuth sign-in to a Tavily account. Unlike Exa and Firecrawl, Tavily has no keyless mode."}],"archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/terraform","slug":"terraform","name":"Terraform","title":"Terraform MCP server: registry docs and HCP Terraform for AI agents","summary":"The Terraform MCP server gives AI agents current Terraform Registry data on providers, modules and policies, and with a token manages HCP Terraform and Terraform Enterprise workspaces, runs and plans.","category":"cloud-and-devops","publisher":"HashiCorp","publisherUrl":"https://www.hashicorp.com","official":true,"repo":"hashicorp/terraform-mcp-server","stars":1540,"repoLicense":"MPL-2.0","lastPush":"2026-10-01","registryName":"io.github.hashicorp/terraform-mcp-server","freeTier":"free-plan","cost":"Free and open source under MPL-2.0; public registry lookups need no account, and the HCP Terraform and Terraform Enterprise tools need an account HashiCorp prices separately.","pricingUrl":"https://github.com/hashicorp/terraform-mcp-server","clients":["VS Code","Cursor","Claude Desktop","Claude Code","Codex CLI","Gemini CLI","Amazon Q Developer","Kiro CLI","Bob IDE and Shell"],"tools":["search_providers","get_provider_details","list_workspaces","get_plan_details","get_plan_logs","get_plan_json_output"],"whatItDoes":["The Terraform MCP server is HashiCorp's own server for connecting AI agents to Terraform data. Its default toolset looks up the public Terraform Registry: providers and their resource docs, modules and policies, so the agent writes HCL against the current provider version instead of guessing at arguments that changed.","With an HCP Terraform or Terraform Enterprise token set in TFE_TOKEN, the Terraform server adds tools for your organization: list and manage workspaces, their variables and tags, read the private registry, and inspect runs. get_plan_details, get_plan_logs and get_plan_json_output let the agent read what a plan will change and why it failed. Tools that need explicit approval stay off unless ENABLE_TF_OPERATIONS is set to true.","HashiCorp ships the Terraform server as a Docker image and a Go binary. It runs over stdio by default or over Streamable HTTP for a shared deployment, with per-session rate limits, CORS rules and an organization allowlist. The README warns that the server may expose Terraform data to the LLM and should not be used with untrusted clients. The repo had 1,539 GitHub stars on 2026-10-01."],"useItWhen":["You want the agent to write a module using the current version of a provider's resources.","You need to compare community modules in the registry for a job.","A plan failed in HCP Terraform and you want the agent to read its logs.","You want to know which workspaces use a variable or tag."],"notFor":["Teams on plain open-source Terraform with local state only: the workspace and plan tools need HCP Terraform or Terraform Enterprise.","Running applies unattended: with ENABLE_TF_OPERATIONS on and a broad token, the agent can change workspaces and start runs, so keep tokens scoped."],"requirements":["Docker, or the terraform-mcp-server binary","An HCP Terraform or Terraform Enterprise API token, only for the organization tools"],"connect":[{"label":"Claude Code, local Docker server","command":"claude mcp add terraform -s user -t stdio -- docker run -i --rm hashicorp/terraform-mcp-server"},{"label":"Codex, local Docker server","command":"codex mcp add terraform -- docker run -i --rm hashicorp/terraform-mcp-server"},{"label":"Gemini CLI extension","command":"gemini extensions install https://github.com/hashicorp/terraform-mcp-server"}],"connectNote":"For HCP Terraform or Terraform Enterprise tools, add -e TFE_TOKEN and -e TFE_ADDRESS to the docker run command, with the token from your HCP Terraform user settings and the address including https. Choose toolsets with --toolsets, for example registry,terraform.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/unity","slug":"unity","name":"Unity","title":"Unity MCP server: control the Unity Editor with AI agents","summary":"The Unity MCP server, MCP for Unity, gives AI agents control of the Unity Editor: create scenes and GameObjects, edit C# scripts, manage assets, run tests, profile and build.","category":"game-and-3d","publisher":"CoplayDev (Aura)","publisherUrl":"https://github.com/CoplayDev","official":false,"repo":"CoplayDev/unity-mcp","stars":14638,"repoLicense":"MIT","lastPush":"2026-09-30","freeTier":"free","cost":"Free and open source under MIT; Unity's editor licensing is separate, and the maintainers' paid Aura assistant is not required.","pricingUrl":"https://github.com/CoplayDev/unity-mcp","clients":["Claude Desktop","Claude Code","Codex","Cursor","VS Code","Windsurf","Cline","Gemini CLI","local LLMs"],"whatItDoes":["MCP for Unity is a community MCP server that connects AI agents to a running Unity Editor. In plain language an agent can create scenes and GameObjects, add components, edit C# scripts, manage assets, run tests, profile and build the project. The README counts 47 tool entrypoints, grouped so you can turn on areas such as VFX, animation, UI and testing only when needed.","MCP for Unity installs as a Unity package from a git URL or OpenUPM, and its editor window configures every detected MCP client in one click. The server part runs on Python 3.10 or newer through uv. It can route to several open Unity instances, validate scripts with Roslyn, and run as a remote server with authentication. Version 10.0.0 shipped on 2026-06-30.","MCP for Unity is sponsored and maintained by Aura, which makes a paid AI assistant for Unity and Unreal, and it states it is not affiliated with Unity Technologies. The repo had 14,633 GitHub stars on 2026-10-01 and was pushed on 2026-09-30. The same team now also makes an open-source Godot server, Godot AI."],"useItWhen":["You want to describe a scene and have the agent build it in the editor.","You want the agent to write a C# script, attach it and test it.","You want to run the test suite and have the agent fix what fails.","You need repetitive asset or prefab changes done across a project."],"notFor":["Projects without version control: the agent edits scripts, scenes and assets directly, so commit before a session.","Older Unity versions: it supports Unity 2021.3 LTS to 6.x.","Official support: it is a community project, not made by Unity Technologies."],"requirements":["Unity 2021.3 LTS to 6.x","Python 3.10 or newer through uv"],"connect":[{"label":"Unity Package Manager, Add package from git URL","command":"https://github.com/CoplayDev/unity-mcp.git?path=/MCPForUnity#main"},{"label":"OpenUPM","command":"openupm add com.coplaydev.unity-mcp"}],"connectNote":"After installing the package, open Window, MCP for Unity, and click Configure All Detected Clients to write the MCP config for each client on your machine. Pin #v10.0.0 instead of #main in the git URL to stay on a release.","archived":false,"checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"},{"url":"https://cookiesforai.app/mcp/vercel","slug":"vercel","name":"Vercel","title":"Vercel MCP server: deployments, logs and projects for AI agents","summary":"The Vercel MCP server gives AI agents access to your Vercel account through Vercel's hosted endpoint: search the docs, inspect projects and deployments, read runtime logs and errors, and query Web Analytics.","category":"cloud-and-devops","publisher":"Vercel","publisherUrl":"https://vercel.com","official":true,"registryName":"com.vercel/vercel-mcp","remoteUrl":"https://mcp.vercel.com","freeTier":"free-plan","cost":"Vercel's docs list no separate charge for the server, which is in public beta; anything it deploys or buys is billed under your Vercel plan.","pricingUrl":"https://vercel.com/docs/agent-resources/vercel-mcp","clients":["Claude Code","Claude.ai","Claude Desktop","ChatGPT","Codex CLI","Cursor","VS Code with Copilot","Devin","Raycast","Goose","Windsurf","Gemini Code Assist","Gemini CLI","OpenCode","Pi"],"tools":["list_deployments","get_deployment","web_fetch_vercel_url","get_runtime_logs","get_runtime_errors","list_projects","get_project","list_teams"],"whatItDoes":["The Vercel MCP server is Vercel's own hosted server for connecting AI tools to a Vercel account, in public beta on 2026-10-01. After OAuth sign-in and granting access to a team, an agent can list projects and deployments, inspect a deployment's status, read and filter runtime logs, group runtime errors by route, fetch protected preview URLs you can access, and query visitors, page views and custom events. Docs search works without signing in.","Vercel's tools reference listed 213 tools in 28 categories on 2026-10-01, well beyond reading: domains and DNS, environment variables, feature flags, firewall rules, rolling releases, sandboxes, and billing and purchases. Vercel's changelog says the server can now deploy code and make purchases, so a connected agent can change production settings and spend money within your permissions.","Vercel only allows AI clients it has reviewed to connect, and requires explicit consent for each client connection to block confused deputy attacks. Its docs list 14 supported clients, and the add-mcp command also configures others such as Cline, Kilo Code and Zed. The server has no public source repo; it is listed in the MCP Registry as com.vercel/vercel-mcp."],"useItWhen":["A deployment failed and you want the agent to read the build and runtime logs.","You want to know which routes are throwing errors in production.","You want the agent to check a protected preview URL after a push.","You want traffic numbers for a page without opening the dashboard."],"notFor":["Agents without confirmation steps: the server has the same access as your Vercel account, including deploys, environment variables and purchases, so keep human approval on.","Unlisted clients: only clients Vercel has reviewed can connect."],"requirements":["A Vercel account, signed in through OAuth with access granted to the team you want","A supported MCP client"],"connect":[{"label":"All detected agents at once","command":"npx add-mcp https://mcp.vercel.com"},{"label":"Claude Code (then type /mcp to sign in)","command":"claude mcp add --transport http vercel https://mcp.vercel.com"},{"label":"Codex CLI","command":"codex mcp add vercel --url https://mcp.vercel.com"},{"label":"Cursor .cursor/mcp.json","command":"{\"mcpServers\": {\"vercel\": {\"url\": \"https://mcp.vercel.com\"}}}"}],"connectNote":"Adding the server does not finish setup: sign in to Vercel in the browser and grant the connection access to each team whose projects the agent should see. Claude.ai and ChatGPT add it as a custom connector with the URL https://mcp.vercel.com, on plans that allow custom connectors.","checked":"2026-10-02","checkedAt":"2026-10-02T03:02:16+02:00","changedAt":"2026-10-03T10:04:47+02:00"}]}