docs(projects): add TeslaMate LLM Bridge to community projects#5409
docs(projects): add TeslaMate LLM Bridge to community projects#5409jakezai250808-sudo wants to merge 1 commit into
Conversation
✅ Deploy Preview for teslamate ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
|
Hopefully database permissions are setup to only allow access to the public schema, and completely deny access to the private schema with the tokens. Sure there may be other safeguards, but regardless I think this one is important and relatively easy. But I don't see any instructions on how to do that.... |
|
Hi @brianmay — thanks, that's exactly the right instinct, and we've now documented it. The bridge only issues read-only SELECTs (every query is a declarative "play" whose SQL is statically validated to reject writes), so it never needs write access or the tokens table. We've added a ready-to-run script (
The install docs (README + both guides) now recommend this restricted role instead of the default full-privilege Done in jakezai250808-sudo/teslamate-llm-bridge#32 / jakezai250808-sudo/teslamate-llm-bridge#33. Thanks for pushing on this — real improvement, and easy as you said. |
|
I have read the CLA Document and I hereby sign the CLA |
Adds TeslaMate LLM Bridge to the community projects list.
What it is
An open-source bridge that exposes a TeslaMate Postgres database to ChatGPT, Claude Desktop, and Coze through declarative-YAML "plays". Each play is one read-only SQL query plus a compute pipeline plus a creative-prompt template, and the same play definition is rendered as a ChatGPT Custom GPT Action, a Coze plugin tool, and an MCP tool — write once, every assistant can call it.
13 plays ship out of the box (driving personality, monthly wrapped, night-owl, early-bird, cumulative climb, charging habits, ...) and contributing a new one is YAML-only, no Java or Python required.
Why submit it here
It is built specifically on top of TeslaMate (reads the standard schema: drives, charging_processes, positions, ...) and the README directs users to install TeslaMate first. Closest existing entries on this page are TeslaMate Achievements (gamified badges) and the various Grafana/HA dashboards — this is a complementary "talk to your car via LLM" angle.
Safety
Plays are not raw SQL access. Each manifest passes a JSON-Schema gate, an SQL static guard (SELECT-only, no DDL/DML, statement timeout, read-only transaction), and an arithmetic-only expression language before being loaded.
Try it without a Tesla
License: AGPL-3.0. Repo: https://github.com/jakezai250808-sudo/teslamate-llm-bridge
Happy to adjust the description if it should be shorter or framed differently. Thanks for maintaining TeslaMate!