{"id":"pinkpixel-dev-mindbridge-mcp","name":"mindbridge-mcp","af_score":48.2,"security_score":39.0,"reliability_score":25.0,"what_it_does":"mindbridge-mcp is a Model Context Protocol (MCP) server/bridge that routes tool calls to multiple LLM providers (e.g., OpenAI, Anthropic, DeepSeek, Google, OpenRouter, Ollama) and exposes MCP tools such as getSecondOpinion plus listing tools like listProviders and listReasoningModels. It also supports an OpenAI-compatible API layer according to the README.","best_when":"You want an MCP tool surface (and optionally OpenAI-compatible compatibility) to unify multiple LLM providers and run multi-model workflows.","avoid_when":"You cannot afford to expose/relay user prompts to multiple upstream provider APIs or you need clear, contract-level guarantees around security controls and reliability/error semantics.","last_evaluated":"2026-03-30T15:29:15.727260+00:00","has_mcp":true,"has_api":true,"auth_methods":["Environment variables for provider API keys (OPENAI_API_KEY, ANTHROPIC_API_KEY, etc.)","Optional API key for OpenAI-compatible providers (OPENAI_COMPATIBLE_API_KEY)"],"has_free_tier":false,"known_gotchas":["Behavior depends on upstream provider responses/limits; routing failures may surface as provider-specific errors.","Since the project is a proxy/router, misconfiguration of provider keys/model names in env/MCP config can lead to tool failures.","No documented guidance here on retries/timeouts/idempotency for multi-call orchestration tools like getSecondOpinion."],"error_quality":0.0}