Help

Last updated 16 October 2026

Model Source finds the AI model your job really needs, not the one everyone says is best. It works inside your AI assistant as a connector (an MCP server).

Connect

• Claude (claude.ai or the desktop app): Settings, Connectors, Add custom connector, and paste https://xfms.xpansion.dev/mcp. Sign in with your email when it opens.
• Claude Code: run claude mcp add model-source --transport http https://xfms.xpansion.dev/mcp, then /mcp to sign in.
• Cursor, Windsurf and other assistants: add an HTTP MCP server at https://xfms.xpansion.dev/mcp.
• Assistants that can't open a sign-in page: sign in here and use the personal connection link from your account page as the server address. Treat it like a password.

Sign-in is free and needs only your email. From the end of Sunday 18 October 2026 it is required.

What you can ask for

• discover: works out what your job really needs, by asking you a few questions.
• rank and pick: the shortlist, or the single best model, for that job.
• benchmark and compare: run the shortlist, or the models you name, on your own example tasks, and judge every requirement. You see the most it can cost first.
• field_results: what has worked on Model Source for jobs of the same kind.
• watch_job: for XF Model Watch subscribers, keeps a job and re-tests it as new models arrive.

Your OpenRouter key

Free comparisons run on your own OpenRouter key. Add it to your assistant's settings for this connector as the X-OpenRouter-Key header. Most assistants do discovery's thinking themselves; if yours can't, that thinking also runs on your key. Model Watch subscribers run on ours, up to their monthly allowance.

More

Daily test results · Privacy · Terms · Questions: russ@visionairy.biz