MCP projects backed by GitHub and Microsoft aim at agentic workflows

Model Context Protocol (MCP) has opened up new ways for AI and agents to interact with tools, codebases, and browsers. Developers are increasingly building on top of the protocol to create browser extensions, editor tools, and other AI-native workflows that are reshaping how code gets written and managed.

To push that work forward, the GitHub Copilot and VS Code teams, in partnership with the Microsoft Open Source Program Office (OSPO), are sponsoring nine open source projects. The funded work clusters around three areas: framework integrations, developer experience tooling, and automation infrastructure for MCP itself.

Framework and platform integrations

The first batch of projects brings MCP into popular frameworks and ecosystems so agents can work with real-world applications and workflows.

  • fastapi_mcp: Exposes secure FastAPI endpoints as MCP tools with minimal setup, authentication, and limited configuration using a unified infrastructure.
  • nuxt-mcp: Nuxt developer tools for route inspection and SSR debugging that help models understand Vite/Nuxt applications.
  • unity-mcp: Interfaces with game engine APIs for AI-assisted game development, letting AI tools manage assets, control scenes, edit scripts, and automate tasks inside Unity.

Developer experience and AI-enhanced coding

A second set of projects focuses on turning LLMs and agents into capable IDE assistants through better workflow ergonomics, deeper code understanding, and safe execution environments.

  • context7: Pulls current, version-specific documentation and code examples and injects them directly into AI and LLM prompt context.
  • serena: Provides semantic retrieval and editing capabilities for agent-driven coding, aimed at improving how agents navigate and modify code.
  • Peekaboo: Swift code analysis that converts on-screen content into actionable AI context, enabling GUI automation for AI assistants.
  • coderunner: Turns LLMs into a local execution partner that writes and runs code in a preconfigured sandbox, auto-installs tools, reads files directly, and returns outputs and generated artifacts.

Automation, testing, and orchestration

The final category covers production-grade tooling for running MCP infrastructure at scale, including testing and debugging support for MCP servers themselves.

  • n8n-mcp: Streamlines workflow creation and orchestration in n8n by integrating AI models to help users understand and work with n8n nodes.
  • inspector: Tests and debugs MCP servers by inspecting protocol handshakes, tools, resources, prompts, and OAuth flows. Includes a built-in LLM playground and eval simulations to catch security or performance regressions.

The projects are among the fastest-growing developer tools in the MCP ecosystem, and the sponsoring teams expect to support further innovation around MCP in the open source community.