MCP server directory

59888 public MCP servers from the official MCP Registry (26001 from verified domains, 33887 community). Search, filter, copy a setup snippet. Nothing here is installed by cosmic-mcp.

Community listing. Not installed or verified by cosmic-mcp. Check the source before you run anything.

Search servers

Search

Showing 1771 to 1800 of 3049

• adrkit decision memory

  dev.adrkit/mcp

  Deterministic, offline, read-only ADR decision memory for coding agents. No model or network calls.

  Verified domain npmstdio memoryai-models

• aeonos

  io.github.Clawdio777/aeonos

  AI search visibility MCP. Get cited by ChatGPT, Perplexity, Claude, and Google AI Overviews.

  Community npmstdio searchai-models

• aether

  io.github.alipishbin77/aether

  Marketplace where AI agents hire other agents, paid per call in USDC. Plus LLM inference.

  Community remotestreamable-http ai-models

• af-databridge

  io.github.farmmer938-lgtm/af-databridge

  Zero-dependency MCP server: CSV to JSON conversion with type inference, and weighted KPI scoring.

  Community ai-modelsspreadsheets

• africa-api-mcp

  io.github.africa-api/africa-api-mcp

  MCP server for Africa API — access data on all 54 African countries through Claude

  Community npmstdio ai-modelsapi-integration

• agent-as-model

  io.github.fernandogarzaaa/agent-as-model

  A true MCP server where the agent is the model and the API key. No API keys.

  Community pypistdio api-integrationai-models

• agent-audit

  io.github.fullstackdegen/agent-audit

  Turn Lighthouse audits into coding-agent fix packs for Claude Code, Cursor, and Copilot.

  Community npmstdio ai-models

• agent-chain-doctor-mcp

  io.github.elfovo/agent-chain-doctor-mcp

  Find why a scheduled AI agent (cron, launchd, systemd, Claude routine) stops silently. Read-only.

  Community ai-modelsautomation

• agent-cost-mcp

  io.github.vk0dev/agent-cost-mcp

  Cost Guard for Claude Code: per-tool ROI, subagent attribution, anomaly/forecast/budget guards.

  Community npmstdio agentsai-models

• agent-eval

  io.github.RudrenduPaul/agent-eval

  Statistical regression testing for LLM agents: p-value, effect size, and CI on behavior change.

  Community pypistdio ai-modelsdevopstesting

• agent-godmode

  io.github.mdvohra/agent-godmode

  Sandboxed workspace MCP: file ops, argv-only run_command, list_files; BYO LLM prompts.

  Community pypistdio ai-modelsfiles

• agent-immune

  io.github.denial-web/agent-immune

  AI agent security: prompt injection detection, semantic memory, output scanning, prompt hardening

  Community pypistdio ai-modelsmemorysecurity

• agent-intercom

  io.github.iambashar/agent-intercom

  Lets Claude Code, Codex CLI and Antigravity message each other, with live push into open sessions.

  Community npmstdio ai-models

• agent-lint

  com.runningunattended/agent-lint

  Checks CLAUDE.md, AGENTS.md or a system prompt for 12 safeguards unattended agents need.

  Verified domain remotestreamable-http ai-models

• agent-magnet

  io.github.helinakdogan/agent-magnet

  Behavioral memory layer for AI tools. Learns preferences and context across Claude, Cursor, Codex.

  Community pypistdio ai-modelsmemory

• agent-mcp-hub

  io.github.blackaxgit/agent-mcp-hub

  One MCP server bridging the Codex, Cursor, OpenCode, Claude, and Antigravity CLI agents

  Community npmstdio ai-models

• agent-media

  io.github.yuvalsuede/agent-media

  AI short drama creator: make micro drama series with consistent casts from Claude, ChatGPT or Codex.

  Community remotestreamable-http ai-models

• agent-output-guard

  io.github.agenson-horrowitz/agent-output-guard

  Validate and verify data from other agents before acting on it. Zero LLM costs.

  Community npmstdio ai-models

• agent-pay-mcp

  com.deepfirstsearch/agent-pay-mcp

  x402 payments in USDC on Base inside an owner-signed, on-chain budget the model can't change

  Verified domain npmstdio ai-modelspayments

• agent-security-scanner-mcp

  io.github.sinewaveai/agent-security-scanner-mcp

  Security layer for AI agents: blocks prompt injection, detects fake packages, scans vulnerabilities.

  Community npmstdio ai-modelssecurity

• agent-skill-loader

  io.github.back1ply/agent-skill-loader

  Dynamically load Claude Code skills into AI agents without copying files.

  Community npmstdio ai-modelsfiles

• agent-toolbelt-mcp

  io.github.marras0914/agent-toolbelt-mcp

  19 focused API tools for AI agents: contract review, prompt optimizer, context packing, and more.

  Community npmstdio ai-modelsapi-integration

• agentaeo-mcp-server

  io.github.digitaldatadriven/agentaeo-mcp-server

  AEO audit tool — measure brand visibility across ChatGPT, Perplexity, Claude, and Google AI.

  Community npmstdio ai-modelssearchsecurity

• agentcast

  io.github.MukundaKatta/agentcast

  Structured-output enforcer: extract and validate JSON from messy LLM text.

  Community npmstdio ai-modelsscraping

• agentcents

  io.github.labham-llc/agentcents

  Local LLM cost proxy: forecasts each call and recommends the cheapest model that clears the task.

  Community pypistdio ai-models

• agentchan

  io.github.damejeras/agentchan

  Lets one coding agent talk to another: Claude Code, Codex or Grok

  Community ai-models

• agentdata-nl

  io.github.DarthSander/agentdata-nl

  x402-paid tools: EU company & sanctions data, web search, LLM chat, crypto intel, x402 monitoring.

  Community remotestreamable-http ai-modelscommunicationcrypto

• agentdocs

  io.github.dancourse/agentdocs

  An agent-first office suite Claude & ChatGPT read and write over one MCP URL.

  Community remotestreamable-http ai-models

• agentdocs

  io.github.titov-d/agentdocs

  Fresh, source-verified docs for building MCP servers and Claude agents, for AI coding agents.

  Community npmstdio ai-models

• agentfenster

  com.agentfenster/agentfenster

  Windows GUI automation on a hidden second desktop, driven from Claude Code over MCP.

  Verified domain pypistdio ai-modelsautomation

Source: the official MCP Registry (registry.modelcontextprotocol.io), refreshed daily, plus public GitHub data. "Verified domain" means the publisher proved the namespace's domain to the registry; it is not a review. Machine-readable: /v1/directory/servers.