{"id":"cognitive-stack-mcphub","name":"mcphub","af_score":49.0,"security_score":35.5,"reliability_score":26.2,"what_it_does":"mcphub is a Python package/CLI that helps developers configure, install/run, and connect MCP (Model Context Protocol) servers into AI applications. It supports stdio-based MCP connections and optionally an SSE-based mode (via a `mcphub run ... --sse` supergateway-style setup), provides a JSON `.mcphub.json` configuration format, and includes adapters to integrate MCP tools with frameworks such as OpenAI Agents, LangChain, and Autogen.","best_when":"You want a developer-focused local/embedded MCP integration layer that orchestrates MCP servers and converts MCP tools for popular agent frameworks, and you control the MCP server sources and runtime environment.","avoid_when":"You need a stable, standardized HTTP API with documented SLAs, authz/authorization per request, or strict operational guarantees from mcphub itself.","last_evaluated":"2026-03-30T13:49:47.106408+00:00","has_mcp":true,"has_api":false,"auth_methods":["Environment variables for required third-party API keys (e.g., OPENAI_API_KEY)","No explicit user-facing auth for mcphub itself is described; server execution/connection is local/subprocess/SSE mode"],"has_free_tier":false,"known_gotchas":["mcphub orchestrates third-party MCP servers by cloning/running commands; agent behavior may depend on external server correctness and startup timing","Auto-configuration from GitHub explicitly uses OpenAI to analyze README; this introduces dependency on external network/API availability","SSE mode exposes endpoints, but no auth/rate-limit behavior is documented in the provided material—agents should not assume safe multi-user exposure without adding safeguards"],"error_quality":0.0}