Development aids
Tools in the repository that make working on FunkArr easier.
In the running app
- Ruleset debugger. Rulesets → + New in the web UI runs your rules against live Mediathek data and shows which entries match and why. It uses
TestScoringWorker, so it scores exactly like a real search. - Scoring history. Every scoring decision is recorded; the history of a ruleset (Rulesets → ruleset → History) shows which rule matched or rejected an entry.
- Live logs. Settings streams the server log, so you don't need
docker logsfor a quick look. - Metrics.
/metricsexposes the Prometheus metrics of every domain, see Observability.
Dev builds
Add the dev-build label to a pull request to publish a container image of it to ghcr.io/st0o0/funkarr. Useful for testing a branch on a real server without a release.
Architecture model
pnpm --dir docs exec likec4 serve architecture opens the model in the LikeC4 viewer with every view. The VS Code extension adds completion and a live preview. See Architecture.
AI agents
The repository is set up for coding agents, but you don't need one to contribute.
AGENTS.mdholds the compact rules,CLAUDE.mdadds the OpenSpec workflow and which skill to load for which task..claude/skills/has templates for recurring work:funkarr-actor,funkarr-message,funkarr-endpoint,funkarr-test, the OpenSpec skills, ande2e-verify, which runs the manual E2E plan against the dev stack.- Agents follow the same OpenSpec flow as people, and their pull requests get the same review.