Back to products
mcp-use

mcp-use

Open source SDK and infra for MCP servers & agents

Overview

What it is

mcp-use is the open-source devtools and cloud infrastructure to help dev teams quickly build and deploy custom AI agents with MCP servers. Our SDK has over 5,000 GitHub stars, 100k downloads, and is trusted by engineers at NASA, Cisco, NVIDIA, etc...

Intent

I need it when

Monitor MCP app usage and catch regressions in production with observability

mcp-use Cloud provides built-in analytics, session replay, error tracking, and alerts. Developers see traffic, tool-call volume, latency, and can replay user conversations end-to-end to identify issues before users report them.

Build and deploy MCP servers and ChatGPT/Claude apps quickly without managing infrastructure

mcp-use provides a fullstack MCP framework with one-command deployment to Manufact Cloud, auto-scaling, observability, and branch previews. Developers scaffold complete MCP servers and React-based apps in under 60 seconds and deploy via git push, eliminating manual DevOps work.

Connect AI agents to multiple MCP servers and tools with a single unified endpoint

mcp-use's MCPClient and MCPAgent implementations allow developers to aggregate multiple MCP servers through one endpoint, supporting OpenAI, Anthropic, Google, and LangChain LLMs. This enables agents to discover and execute tools across distributed servers without friction.

Publish MCP apps to ChatGPT and Claude marketplaces with minimal friction

mcp-use automates marketplace submission by generating logos, copy, and screenshots. It provides submission checklists and embedded chat widgets, reducing the barrier to reaching users across ChatGPT, Claude, and other MCP clients from a single codebase.

Test and debug MCP servers against real LLM clients before production deployment

The built-in Cloud Inspector lets developers test MCP servers against GPT, Claude, and Gemini models, swap models mid-session, inspect JSON-RPC calls, and run automatic evaluations—all from a browser without local setup.

Drop

Not a fit when

  • User needs a closed-source, proprietary MCP solution without open-source components
  • User requires MCP infrastructure that does not support TypeScript or Python development
  • User needs unlimited analytics retention and premium support on the free tier
  • User wants to avoid GitHub-based deployment workflows and CI/CD integration
  • User requires on-premises deployment without cloud hosting options
Commercials

Pricing

USD0 - USD250 / monthly View pricing