MCP server directory
55,431 MCP servers that connect AI assistants to tools and data, from the official MCP Registry and the community lists, updated every hour.
- 16 categories
- 35,580 remote servers
- Updated every hour
com.mspmcp/instinctcom.mspmcp/instinct
MSP market intelligence for AI assistants: profiles, tech, maturity scores, and change signals.
io.github.yusufcemres/uisightio.github.yusufcemres/uisight
Your AI sees the screen but can't measure it — measured web UI findings, shared live sessions
Glint Meet — every global model for your agentio.github.zzzjy765/glint-meet
Call every global AI model through one OpenAI-compatible API. One key, all models.
io.github.LBZ558/prophetkey-mcpio.github.LBZ558/prophetkey-mcp
ProphetKey MCP bridge - encrypted local-first API key vault for AI agents, keys never exposed
Revnuvo Company Intelligenceio.github.hardeyhemy/revnuvo-company-intelligence
Observed company technologies, changes and signals for AI agents, with evidence and confidence.
元铸工坊 SOSOFAST (yuanzhu)io.github.RavenWangChina/yuanzhu
Governed AI workflow assets: staged writes, evals-gated template forging, ontology via MCP.
com.super-log/super-logcom.super-log/super-log
Query your local super-log bench from an AI agent - read-only log tools over the hub.
com.milotche/carre-de-sablecom.milotche/carre-de-sable
Un lieu réservé aux agents IA. Une consigne par jour. Ce que vous y déposez reste.
io.github.dkaleganov/mercury-multiorg-mcpio.github.dkaleganov/mercury-multiorg-mcp
Unofficial read-only MCP server exposing several Mercury banking organizations to one AI session.
io.github.zapsoblige-hash/policyvaultio.github.zapsoblige-hash/policyvault
Non-custodial Kaspa policy enforcement for AI agents; signers retain custody.
com.moosejudge/castor-mcpcom.moosejudge/castor-mcp
Castor for Confluence Data Center — MCP server bridging Confluence DC with AI agents
Intestaio.intesta/registry
Trust registry for agents: check_trust verdict A0-A4, signed fact passports, honest refusals.
io.github.BotHireAgent/bothireio.github.BotHireAgent/bothire
Hire, pay gaslessly, deliver and get paid between AI agents on BotHire (USDT/USDC, escrow).
com.edgethirteen.www/killcordcom.edgethirteen.www/killcord
Dead-man's switch for Anthropic API spend: kills the key the instant it crosses a limit you set.
io.github.linxule/vellumio.github.linxule/vellum
Shared living surface where AI agents leave short traces and weave lineages
com.edgethirteen.www/apidriftcom.edgethirteen.www/apidrift
Alerts on breaking changes, deprecations, and pricing shifts across major LLM provider APIs.
app.hykeep/keepapp.hykeep/keep
Keep — AI-first delivery PMO: projects, records, tasks, cycles, requirements, artifacts and time.
VoxelDraft – AI Voxel Modelingio.github.huruki-geo/voxeldraft
Create, edit, render, save, and export voxel models with interactive 3D previews.
com.edgethirteen.www/kdpcheckcom.edgethirteen.www/kdpcheck
Scan manuscript text for AI-slop prose patterns before publishing to Amazon KDP.
io.github.eliottreich/taskbounty-mcp-serverio.github.eliottreich/taskbounty-mcp-server
Hire agents, request quotes, delegate funded tasks, submit work, and collaborate on missions.
io.github.yantrikos/yantrikdb-mcpio.github.yantrikos/yantrikdb-mcp
Cognitive memory for AI agents — semantic recall, knowledge graph, and contradiction detection
io.github.parkschun79/diversz-commonsio.github.parkschun79/diversz-commons
MCP tool window for autonomous AI: principles, join, talk, treasury, signal.
com.edgethirteen.www/cloakcheckcom.edgethirteen.www/cloakcheck
Scan a page for content planted to hijack an AI browsing/shopping agent before it acts on the page.
io.github.jdhart81/wu-wei-routerio.github.jdhart81/wu-wei-router
Plan agent workload routes from evaluation and cost data. $1 USDC per plan; modeled savings only.
io.github.priyankark/hci-quality-mcpio.github.priyankark/hci-quality-mcp
Interface observations and explicit GOMS/KLM, Fitts, Hick–Hyman and cognitive walkthrough models.
io.github.jdhart81/maxwell-defenseio.github.jdhart81/maxwell-defense
Model proof-of-work policy and latency for AI agents. $1 USDC rehearsal; no runtime protection.
EventEdge Oraclespace.arakelproof.edge/eventedge
Prediction-market monitoring, AI-agent context, and recurring x402 machine checks.
io.github.iamalanlui/apipayio.github.iamalanlui/apipay
Prepaid balance for AI agents: one key, 4,900+ tools your agent can run today, caps, receipts.
io.github.vnbochkarev-netizen/vibo-mcpio.github.vnbochkarev-netizen/vibo-mcp
Persistent memory for AI agents: L1/L2/L3, web compression, threads.
io.github.raclarke1/hivepostsio.github.raclarke1/hiveposts
Agent library of verified fixes under permanent URLs.
Revensicom.revensi/revensi
Connect your AI assistant to Revensi OS agents, workflows, and business data.
com.failecho/failechocom.failecho/failecho
Check what other agents hit the same tool failure — and what recovery worked. Ask before retrying.
app.vercel.cloakcheck-wheat/cloakcheckapp.vercel.cloakcheck-wheat/cloakcheck
Scan a page for hidden prompt-injection payloads targeting AI agents.
io.github.0rkz/byte-protocolio.github.0rkz/byte-protocol
PayPerByte — per-byte data for AI agents: x402 USDC on Base, EIP-712-attested. No token.
io.github.Jason26214/omni-computer-useio.github.Jason26214/omni-computer-use
Windows MCP server replicating Anthropic's computer-use tool surface, for any MCP client.
io.github.JoaoBerne/cli-bridge-mcpio.github.JoaoBerne/cli-bridge-mcp
Other AI CLIs (Gemini, GPT, Claude, opencode, Ollama) as a council — ban-safe, no API keys.
ai.oddsockets/remote-mcpai.oddsockets/remote-mcp
Agent-first real-time messaging over MCP: self-provision channels, publish, history, presence.
io.github.guildbook-ai/guildbook-mcpio.github.guildbook-ai/guildbook-mcp
Work-history network for AI agents: register, find agents by skill, consult a free house C-suite.
io.github.Fluenta-Team/fluentaio.github.Fluenta-Team/fluenta
Score business ideas on six live market signals. Launch Readiness Score (LRS) for AI agents.
io.github.cmendezs/mcp-nfe-brio.github.cmendezs/mcp-nfe-br
MCP server for Brazilian e-invoicing: NF-e and NFC-e (modelo 55/65), schema 4.00, SEFAZ.
Agent Tavernio.github.tima100faces/agent-tavern
Public board where AI agents ask, answer and post findings across runtimes.
DNSMintcom.dnsmint/mcp
Mint a secure HTTPS endpoint for your agents. Each on its own dedicated domain.
Federal Contract Lifecycle Evidenceio.github.ThryveBaseline/agent-services-003
Resolve a federal award and buy bounded USAspending contract lifecycle and end-date evidence.
dev.glbforge/glbforgedev.glbforge/glbforge
Make AI-generated 3D assets web-ready: budgets, SSIM-verified optimization, STL/USDZ export.
com.marutap/marutapcom.marutap/marutap
Agent requests a purchase, the human taps approve, the agent gets a card scoped to it.
io.github.MaterialModel/materialmodelio.github.MaterialModel/materialmodel
Explore discoveries, ask questions, share findings, and build collaborations with other agents.
io.github.andrewnexa-hub/guildbook-mcpio.github.andrewnexa-hub/guildbook-mcp
Work-history network for AI agents: register, find agents by skill, consult a free house C-suite.
com.ratchetgate/ratchetcom.ratchetgate/ratchet
Your agent asks before it acts, so the same real-world action is attempted at most once.
From the official MCP Registry. Listing is not an endorsement: review a server's code and permissions before giving it access to your data. Last synced Oct 31, 2026.
Frequently asked questions
What is MCP?
The Model Context Protocol is an open standard that lets an AI assistant (Claude, ChatGPT, Cursor, an agent) call outside tools and read outside data through a "server". A remote server is a URL you add to your client; a local one runs on your computer from a package (npm, PyPI, Docker).
How do I add an MCP server to Claude or Cursor?
Open the server's page here and copy its configuration block into your client's MCP settings file, then restart the client. Remote servers need only the URL; local ones need Node (npx), Python (uvx) or Docker installed.
Are these servers safe?
Anyone can publish to the registry. A server can read what you give it access to and act on your behalf, so prefer servers with public source code from a publisher you recognise, and grant the smallest scope that works.