Skip to content
Browse documentation

Integrating Cursor

Cirvix rewrites Cursor's MCP configuration so existing servers route through the gateway, without changing how you invoke them.

Cursor's MCP configuration lives at ~/.cursor/mcp.json. The gateway reads that file verbatim as its upstream list, so the servers you already have route through Cirvix without a parallel configuration to keep in sync.

Configure

TerminalShell
$ cp ~/.cursor/mcp.json ~/.cirvix/upstream.json $ cirvix gateway --servers ~/.cirvix/upstream.json   [cirvix] gateway up · 2 upstream · 9 rules · local policy

Then replace ~/.cursor/mcp.json with a single entry naming the gateway. An entry called cirvix in the upstream file is skipped, so the copy is safe to make after the switch as well as before.

Nothing else changes

The tool names, arguments, and results Cursor sees are identical — the only difference is that each call is now decoded, checked against policy, and recorded before it reaches github-mcp or postgres-mcp.

Note
Current integration status is tracked on the Integrations page.
Edit this page on GitHub
Was this page helpful?