MCP in Eclipse: connect AI agents to your tools

The Model Context Protocol (MCP) is an open standard that lets AI agents do more than chat — they can call real tools, read your files, run Git actions and reach external data. Eclipse speaks MCP out of the box.

What MCP unlocks

Connect your model to your tools through one open protocol instead of one-off plugins.

Act on your code

Agents can open, edit and refactor files and run Git operations — with your approval — instead of only suggesting text.

Reach external systems

Connect MCP servers for issue trackers, databases, docs and internal APIs so the assistant works with your real context.

Open & portable

MCP is a shared standard, so the tools you connect aren't locked to one vendor or one model.

Bring your own tools

Add MCP servers for the systems your team already relies on, and distribute them the way you distribute other extensions. Agents in Eclipse then call those tools as part of their workflow — from committing code to querying a service.

Pair MCP with any model →
# enable an MCP tool server
mcp.servers:
  - name: "git"
  - name: "filesystem"
  - name: "your-internal-api"
 
# agents can now call these tools

About MCP

Does Eclipse support the Model Context Protocol?

Yes. Eclipse supports MCP, so AI agents can connect to external tools, services and data through a single open standard instead of bespoke integrations.

Can I add my own MCP servers?

Yes. You can connect MCP servers for your own tools and data, and distribute them like other extensions, so agents in Eclipse can act on the systems your team already uses.

Does the agent run tools without my approval?

No. You stay in control — Eclipse asks for your approval before an agent runs actions like editing files or executing Git commands through MCP.

Give your AI real tools

Download Eclipse free and build agentic workflows on an open protocol.

Download free for Windows