Model Context Protocol (MCP) Server Directory

Enterprise teams store 80% of their institutional knowledge in Notion—but AI agents can't access it. This FastMCP TypeScript server exposes Notion pages, databases, and wikis to Claude Desktop and Cursor agents with semantic search, auto-summarization, cross-database joins, and incremental indexing that keeps knowledge fresh without API rate limit exhaustion.

Cerebras CS-4 delivers 30x faster inference than NVIDIA Blackwell by processing on a full 300mm wafer. This FastMCP server wraps Cerebras' API, giving Claude Desktop and Cursor agents sub-100ms token generation for latency-critical tool calls.

SambaNova's SN50 Reconfigurable Dataflow Unit runs any model — from 1B to 405B parameters — on a single chip with automatic model switching. This FastMCP server wraps SambaNova's API for enterprise agents needing multi-model inference with SLA guarantees.

Groq's Language Processing Units deliver deterministic inference with zero GPU contention. This FastMCP server wraps Groq's API for Claude Desktop and Cursor agents, achieving consistent sub-50ms TTFT at $0.05/M tokens — the cheapest ultrafast inference available.

Apple M5 Ultra with 512GB unified memory enables running 400B models locally. This FastMCP server exposes Core ML and MLX inference as MCP tools, giving AI agents on-device inference without cloud API costs.

Firecrawl is the #1 MCP server for web context in 2026, used by thousands of developers for search, scrape, parse, crawl, and interact operations. This FastMCP server wraps Firecrawl's capabilities for AI agents needing real-time web intelligence.

Skild AI's S1 foundation model learns robot tasks from single videos. This FastMCP server exposes S1's capabilities as MCP tools, enabling AI agents to teach robots new tasks, monitor execution, and coordinate multi-robot fleets.

OpenAI's Assistants API sunset on August 26, 2026, leaving 2.3M API keys stranded. This FastMCP server translates legacy Assistants API calls to the new Responses API format, automatically converting threads, runs, and function calls to MCP-compatible tool invocations.

Claude Opus 5 costs $5/$25 per million input/output tokens — the same as Opus 4.x but approaching Fable 5 capabilities at half the cost. This FastMCP server tracks real-time token spend, enforces per-session budget gates, and auto-routes to Sonnet 5 when thresholds are breached.

Z.ai's GLM-5.3-Flash went viral within 48 hours of its August 26 release — a 320B-A18B natively multimodal MoE under MIT license at $0.075/M input. This FastMCP server exposes its text, image, and video capabilities as MCP tools for Claude Desktop and Cursor.

Distributed tracing data lives in Jaeger and Grafana Tempo dashboards that agents can't query. This FastMCP server bridges the gap, exposing OpenTelemetry traces, service dependencies, and latency percentiles directly to AI agents in Claude Desktop and Cursor for real-time debugging.

Terraform state files hold the complete picture of your infrastructure, but querying them requires manual JSON parsing. This FastMCP server exposes Terraform state intelligence — resource inventory, drift detection, and cost estimation — directly to AI agents in Claude Desktop and Cursor.