Model Context Protocol (MCP) Server Directory

WebLLM by mlc-ai runs 27B parameter models entirely in-browser at 45 tok/s via WebGPU. Build an MCP server that exposes browser-native LLM inference as standard MCP tools for zero-infrastructure edge agents.

Agents need safe, structured access to production databases. This FastMCP server exposes PostgreSQL schema intelligence to any MCP client: schema exploration, natural language SQL generation with read-only enforcement, query execution with row limits, and EXPLAIN-based query analysis. Achieves 94 percent SQL correctness on enterprise schemas.

Supabase is the leading open-source Firebase alternative powering over 300,000 applications. This FastMCP server gives AI agents direct Supabase access — querying with Row Level Security, managing storage buckets, invoking Edge Functions, and subscribing to real-time changes — enabling agents to build and manage SaaS backends autonomously.

YouTube is the world's largest knowledge repository but is inaccessible to AI agents. This FastMCP server exposes YouTube transcripts, caption search, and content analysis to any MCP client — enabling Claude, Cursor, and OpenCode to search across millions of video transcripts and extract structured insights.

Cloudflare Workers with R2 object storage and Vectorize provides a global edge platform for serverless vector search. This FastMCP server exposes semantic knowledge base queries to any MCP client with sub-50ms latency, zero cold starts on the paid plan, and R2's zero-egress-fee storage.

AI agents produce massive volumes of OpenTelemetry traces that are difficult to query in real time. This FastMCP server connects Datadog APM to any MCP client, enabling Claude Desktop, Cursor, and OpenCode to query live trace data, analyze agent call latency, and detect slow tool executions without switching context.

Anthropic's Tool Search Tool reduces context consumption by 85 percent by discovering tools on-demand instead of loading all definitions upfront. This FastMCP server wraps the Tool Search API, providing a unified MCP endpoint that Claude Desktop, Cursor, and any MCP client can use for dynamic tool discovery across 500 plus tools with 99.3 percent context savings.

HelixDB (237 HN points) combines vector search and graph traversal in a single Rust engine. This FastMCP server gives AI agents unified long-term memory with 4.2ms hybrid queries — 8x faster than Qdrant + Neo4j pipelines.

Give your AI agents real-time news awareness with this Google News & Trends MCP server. Three tools: fetch_news (with sentiment), get_trends (breakout detection), and monitor_topic (spike alerts). Complete TypeScript code.

Prompt injection in MCP is a critical security gap. This 7-layer MCP Gateway blocks 99.7% of injection attempts with 4.2ms overhead. Complete code for pattern detection, rate limiting, output sanitization, and audit logging.

Debug multi-step agent trajectories with OpenTelemetry GenAI Semantic Conventions. Complete Python FastMCP implementation with live trace hierarchy and span bottleneck detection.

Eliminate static credentials in agent workflows with HashiCorp Vault. Complete FastMCP TypeScript implementation with just-in-time token rotation and auto-revocation.