Deploy the worker
Run the container with Docker Compose or the published Helm chart. One image serves the worker, setup flow, and docs.
Echolocation for your merge requests.
ReviewPhin
Self-hosted, agentic code review for GitLab and GitHub. Let engineers spend review time on design, risk, and intent instead of hunting typos and obvious errors.
Signal path
Five moves from deploy to comment.
Wire the worker once; it repeats the same path on every trigger.
Register a platform
Store reusable GitLab or GitHub credentials as a platform connection, then attach projects or repositories as tenants.
Choose models
Use the bundled Copilot CLI path or create named model profiles for OpenAI-compatible, Azure, or Anthropic endpoints.
Pick storage
Start with SQLite, switch to Flotiq when an admin panel helps, or load a custom storage adapter.
Publish reviews
Trigger a review from the code review surface and let ReviewPhin reconcile findings, replies, summaries, and state.
Quickstart
Surface it in two environments.
The container image serves the worker, setup flow, and docs from one runtime. Persist /app/data and /app/tmp in production.
# bring the worker onlinedocker compose up -d# confirm it is listeningcurl http://localhost:3000/healthz{"status":"ok"}
REVIEWPHIN_VERSION=1.6.0REVIEWPHIN_CHART=oci://ghcr.io/cdwv/charts/reviewphin# create a secret from runtime envkubectl create namespace reviewphinkubectl create secret generic reviewphin-env \ --namespace reviewphin --from-env-file=.env.production# install onto your clusterhelm upgrade --install reviewphin "${REVIEWPHIN_CHART}" \ --namespace reviewphin --create-namespace \ --version "${REVIEWPHIN_VERSION}" \ --set application.envSecret=reviewphin-env \ --set persistence.size=1Gi
Where it fits
Built for operators who need review automation without giving up deployment control.
- GitLab first. GitLab is the primary setup path, with GitHub supported through an app-based flow.
- Provider-shaped. Platform and storage providers can be replaced without changing review orchestration.
- Container-ready. The image serves the worker, setup assets, and docs from the same runtime.