Adding a technical note that I don't think has come up in this thread.

The v2 API already depends on rules internally: POST /v2/transactions accepts apply_rules: true. So the API will execute your rule set on inbound transactions but provides no way to read, audit or manage the set it just ran. That asymmetry is what makes this feel like a gap rather than a scoping decision.

Two consequences for anyone ingesting from custom sources: rules are the only part of a Lunch Money configuration that can't be exported, version-controlled or restored, and there's no programmatic way to determine which rule categorised a given transaction when something lands unexpectedly.

If full CRUD is too large for the alpha, a read-only GET /v2/rules would cover backup and audit on its own and would be a genuinely useful increment.

I'd love this! I have resorted to reverse-engineering the browser API to manage rules programmatically for now, but it's annoying, and requires managing user cookies.

I would really like to use AI to update my rules through an MCP, just need this endpoint.

Please do this, yes! I'm building something that would really benefit from this (as would other community members!)

Really would love this.

Huge upvote here from me. In the age of LLM's, being able to have an AI go through and figure out merchant information and then also apply rules automatically is a must-have. Without it, we need to instruct AI to run every so often to manually execute rules client-side.

This would be very useful..

Was surprised they did not exist :(