KeelenAnswer · MCP
Can I run a whole project from Claude or another AI chat?
Updated 11 September 2026
Yes, through the Model Context Protocol. MCP lets an AI client call outside tools. So your agent can do more than read context. It can act. Most MCP servers today supply context. That means files, databases, browsers, and issue trackers. A smaller number do work. Keelen's hosted server is the second kind. Register it, say "set me up". Your agent then creates the account, connects your GitHub repository, and submits build requests. GitHub and provider setup can require browser approval. After that you steer Requests from the chat window. Review and merge results depend on the project and repository controls you set.
Register it in Claude Code with one command. Run claude mcp add --transport http keelen https://keelen.ai/mcp
What to do, in order
Pick a client that supports MCP
Claude Code, Claude Desktop, Codex, Cursor, and Windsurf all support it. The protocol is the same in each one. So the server works wherever you already are.
Register the server
For Claude Code that is one command: claude mcp add --transport http keelen https://keelen.ai/mcp. Other clients take the same URL in their own configuration.
Let the agent do the setup
Hand it the setup instructions and say "set me up". It asks for your email. Then it asks for the code that lands in your inbox. From there it guides engine connection, GitHub approval in the browser, and the first project.
Steer in plain language
Submit requests, reorder the roadmap, check status, and read the pull requests that come back, all from the chat window. The web dashboard is optional.
Keep the gates on
Driving from chat changes the interface. The safety model stays as it is. What stands between a request and the main branch? A verified test command, a configured review and required checks.
Context servers and working servers
MCP servers can supply context. They can also act on other systems. Keelen takes development Requests. It shows project records too. Your work can keep going after you close the chat. That depends on capacity, provider limits and the calls that need you.
Follow the setup for your client
Use the public setup guide for your MCP client. Confirm your email. Connect AI credentials in Keelen's dashboard. Approve GitHub access in the browser before you ask the agent to submit work. Keep provider credentials out of the chat.
What happens after you close the chat
The request you submitted becomes roadmap items and tasks ready to build. The loop implements them in isolated VMs. It uses your own credentials and opens one pull request at a time. It keeps going on its own schedule. You can check in later from any client or from the dashboard. You can pause it whenever you like.
The safety model does not change
A separate adversarial reviewer reads the diff. It does not see the conversation that wrote the code. The review model comes from your project settings. Red first proof and clean checkout tests need a verified project test command. Gated auto-merge needs the repository checks and a review window you set up. Plan review, manual merge, and branch-only mode each pick where a person must decide. Check the real evidence before you accept the result.
Where the code runs
Each run gets its own single use VM. It runs as a user without root and is destroyed when the run ends. Network egress goes through a proxy that denies by default. The GitHub token a run holds works on one repository for about an hour. It cannot change your CI workflows.
Build an internal tool and keep it improving
Use Keelen's MCP to submit development requests for an internal tool. You can read the progress there too. A skills or artifact hub can expose its own MCP. That MCP is a feature of the software you build. Its hosting and access rules belong to that project. An outside scheduler can also call Keelen's scan tools to repeat reviews. These are separate links with different jobs.
When Keelen is not the answer
- Your client does not support MCP. The dashboard does everything the server does.
- You want the agent to edit files in your local checkout. That is a job for a CLI coding agent. Keelen works on the repository in its own isolated VMs.
- You are not at ease with a hosted service. It connects to your GitHub repository.
Hand the work to a loop
Connect a repository and write what you want in plain language. Review the tested pull requests that come back.
FAQ
What is MCP?
The Model Context Protocol is an open standard. AI clients use it to call external tools and data sources through one interface. It lets an assistant act on a system. That is more than describing it.
Do I need the web dashboard at all?
You do not need the dashboard. Requests and status work in the chat. GitHub and provider connection can still need browser approval. Use the dashboard to see the roadmap and metrics.
Which MCP clients work with it?
Any client that supports MCP over HTTP. There are setup guides for each client: Claude Code, Claude Desktop, Codex, Cursor, Windsurf, and generic clients.