Discover MCP servers. Connect your agent.
Official Anthropic Claude API integration. Send prompts to Claude models, manage conversations, and access Claude capabilities directly through MCP tools.
https://github.com/anthropics/anthropic-mcp
stdio
api_key
4 tools
llmclaudeanthropicofficialai
Registered 2026-02-17
Official Chroma MCP server. Open-source embedding database for AI applications. Create collections, add documents with embeddings, and perform semantic search queries.
https://github.com/chroma-core/chroma-mcp
stdio
open
8 tools
vector-databaseembeddingsopen-sourceragai
Registered 2026-02-17
Cohere MCP server for enterprise AI. Text generation, reranking, and high-quality embeddings optimized for RAG pipelines. Command and Embed models with multilingual support.
https://github.com/cohere-ai/cohere-mcp
stdio
api_key
5 tools
llmembeddingsrerankingragenterpriseai
Registered 2026-02-17
Official Exa MCP server for AI-powered web search. Semantic search, advanced web search, code context retrieval, website crawling, company research, people search, and deep research capabilities. Neural search engine designed for AI agents. Install via npm: exa-mcp-server
https://github.com/exa-labs/exa-mcp-server
stdio
api_key
8 tools
searchweb-searchsemantic-searchresearchaineural-searchofficial
Registered 2026-02-09 by
Exa
Groq MCP server for ultra-fast LLM inference. Access Llama, Mixtral, and Gemma models with sub-second response times via Groq hardware acceleration.
https://github.com/groq/groq-mcp
stdio
api_key
3 tools
llminferencefastopen-sourceai
Registered 2026-02-17
Official Hugging Face MCP server connecting AI assistants to the Hugging Face Hub. Search models, datasets, and papers. Dynamically connect to Gradio-based tools hosted on Spaces for extended ML capabilities. Supports STDIO, SSE, and Streamable HTTP transports. Open source with 164+ unique weekly MCP clients. Install via npm: @huggingface/mcp-server
https://github.com/huggingface/huggingface.js/tree/main/packages/mcp-server
stdio
api_key
12 tools
aimlmodelsdatasetshuggingfaceinferenceofficial
Intelligent LLM request routing with cost optimization, fallback strategies, and load balancing across 15 models from 7 providers (Anthropic, OpenAI, Google, Meta, Mistral, Cohere, DeepSeek). 6 MCP tools: route_request, get_model_registry, get_routing_strategies, check_model_health, report_model_status, compare_models.
https://llm-router-mcp.fly.dev/mcp/sse
sse
open
6 tools
aillmroutingcost-optimizationload-balancingmodel-selection
Registered 2026-02-09 by
cairn
Mistral AI MCP server. Access Mistral, Mixtral, Codestral, and other models. Chat completions, embeddings, and function calling. European AI alternative with strong reasoning capabilities.
https://github.com/mistralai/client-python
stdio
api_key
4 tools
llmaimodelseuropeanopen-source
Registered 2026-02-17
Official OpenAI Agents SDK with MCP support. Build AI agents that connect to MCP servers for tool integration. Supports both MCP client (consuming tools) and MCP server (exposing tools) patterns. ChatGPT desktop app supports MCP connections. Codex agent framework with native MCP transport. Industry-standard AI platform integration.
https://github.com/openai/openai-agents-python
stdio
api_key
10 tools
aillmagentsopenaichatgptcodexofficial
Registered 2026-02-09 by
OpenAI
Official Pinecone MCP server for vector database operations. Search documentation, create indexes with integrated inference, upsert and search records, cascading search across multiple indexes with deduplication and reranking. Designed for RAG, semantic search, and AI agent memory. Install via npm: @pinecone-database/mcp
https://github.com/pinecone-io/pinecone-mcp
stdio
api_key
8 tools
vector-databaseembeddingssearchragsemantic-searchaiofficial
Official Qdrant MCP server. Semantic search and vector operations for AI applications. Store, search, and retrieve high-dimensional embeddings with filtering and payload support.
https://github.com/qdrant/mcp-server-qdrant
stdio
api_key
5 tools
vector-databaseembeddingssemantic-searchairag
Registered 2026-02-17
Official Replicate remote MCP server for AI model inference. Run thousands of open-source models including image generation, language models, audio, and video. Hosted at mcp.replicate.com with automatic API updates. Supports model search, prediction creation, and result retrieval. No local install needed.
https://mcp.replicate.com
sse
bearer
8 tools
aimlinferencemodelsimage-generationreplicateremoteofficial
AI-powered browser automation built on Playwright. Uses natural language instructions to interact with web pages — click, type, extract data, and navigate without writing selectors.
https://github.com/browserbase/stagehand/tree/main/mcp
stdio
open
5 tools
browserautomationaiplaywright
Registered 2026-02-17
Official Tavily MCP server for real-time web search and data extraction. Four tools: search (AI-powered web search with filtering), extract (content extraction from URLs), map (website structure mapping), and crawl (deep website crawling). Production-ready with configurable depth, domain filtering, and topic specification. Install via npm: tavily-mcp
https://github.com/tavily-ai/tavily-mcp
stdio
api_key
4 tools
searchweb-searchextractioncrawlingresearchaiofficial
Registered 2026-02-09 by
Tavily
Together AI MCP server for open-source model inference. Access Llama, Mistral, Qwen, and 100+ open models. Batch inference, fine-tuning, and dedicated endpoints available.
https://github.com/togethercomputer/mcp-server-together
stdio
api_key
4 tools
llmopen-sourceinferenceaimodels
Registered 2026-02-17
Weaviate MCP server for AI-native vector databases. Semantic and hybrid search, multi-modal retrieval, generative search, and knowledge graph operations. Supports multi-tenancy and real-time data ingestion.
https://github.com/weaviate/mcp-server-weaviate
stdio
api_key
9 tools
vector-databasesemantic-searchaiknowledge-graphrag
Registered 2026-02-17