M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
COMPARE UP TO 4 SERVERS

Compare MCP servers

Compare scores, permissions, risks, and fit in one decision-focused table.

DimensionFastify MCP ServerFastify plugin for building MCP servers with SSE, stdio, OAuth 2.1, and Redis-backed scaling.MartinLoop MCP ServerAn independent control layer that gives AI coding agents budgets, brakes, and receipts
FMRS63 / 100 · C79 / 100 · B
Reliability9 / 2012 / 20
Security and permissions14 / 2018 / 20
Maintenance10 / 2016 / 20
Documentation17 / 2018 / 20
Setup experience13 / 2015 / 20
Best for
  • Node.js/Fastify developers who need to build a scalable, authenticated MCP server from scratch
  • Teams deploying multiple server instances behind a load balancer with Redis for shared state
  • Projects that want TypeBox-based compile-time and runtime type safety for tool inputs
  • Engineering teams scaling AI coding agents while controlling spend
  • Platform teams that need auditable, rollback-aware agent execution records
  • Organizations using multiple agents (Claude, Codex, Gemini) under one governed flow
Not for
  • End users looking for a ready-made server that already exposes a fixed set of tools (e.g. filesystem or database access)
  • Teams not using Node.js/Fastify
  • Simple single-instance prototypes that don't need scaling, OAuth, or session persistence
  • Developers who just want a simple code-completion helper without run governance
  • Teams unwilling to accept local JSONL run records with HMAC signing
  • Environments without a supported adapter (Claude/Codex/Gemini CLIs, direct provider, or verifier-only)
Required permissions
  • Runs as a library inside the developer's own Fastify process — actual filesystem/network access depends entirely on the tool handlers the developer implements
  • Requires network access to a Redis instance when horizontal scaling is enabled (session store and message broker)
  • Requires network access to the configured JWKS or token-introspection endpoint when OAuth 2.1 authorization is enabled
  • URL-mode elicitation sends the end user out of band to an external URL to complete sensitive interactions
  • Read/write access to the configured repo root (writable scope restrictable via --allow-path/--deny-path)
  • Ability to invoke the chosen agent CLI (Claude/Codex/Gemini, etc.) and its underlying model account
  • Local disk writes for run records, receipts, and share outputs (share/run-receipt., etc.)
  • Execution of user-configured verifier commands (e.g. npm test)
Risks and side effects
  • The plugin ships with no built-in tools, so its real-world security depends on the handlers developers write on top of it
  • Without authorization enabled, tasks are protected only by their random UUID — anyone holding a task ID can reach it, since no requestor identity exists to bind it to
  • Production-grade rate limiting must be added by the integrator; the plugin only provides basic DoS protections (length, depth, and property-count limits)
  • Multiple SSE connections are intentionally allowed per session and receive broadcast messages, a deliberate deviation from the MCP spec's single-stream delivery rule that can cause duplicate message delivery
  • `transformRouteSchema` only edits OpenAPI documentation metadata — it does not enforce authentication, so real auth hooks must be configured separately
  • Verifier commands are executed; poorly chosen verifier commands carry security risk despite preflight policy checks
  • Cost and token figures may be estimates (labeled with provenance) and should not be treated as settled accounting
  • Receipt integrity must be verified before use as trustworthy evidence; missing rollback evidence is flagged as EVIDENCE_BOUNDARY
  • Policy defaults come from martin.config.yaml and can be overridden by CLI flags; misconfiguration can loosen safety boundaries
Supported clientsClaude Code, Codex, Gemini CLI, Cursor, VS Code, GitHub Copilot, Continue
Tools00