Aktenlens is a multi-tenant AI document intelligence platform. It is in early development.
- A Go API with two health endpoints:
GET /healthzanswers200when the process is alive.GET /readyzanswers200when the database is reachable, and503when it is not.
- An HTTP server with timeouts and graceful shutdown.
- A PostgreSQL connection pool.
- PostgreSQL with the pgvector extension, running in Docker Compose.
- Database migrations with goose (first migration: the
tenantstable). - A small Docker image for the API (multi-stage build, distroless, runs as a non-root user).
- Automated checks on every pull request: formatting,
go vet, tests with the race detector,golangci-lint(with gosec) andgovulncheck.
- Go 1.26 or newer (needed for the tests and for
make migrate) - Docker
makeand a bash shell (on Windows, use Git Bash)
First, create your local settings. Open .env and replace change-me with a password (letters and digits only):
cp .env.example .env# Start PostgreSQL and the API (press Ctrl+C to stop)
make devIn another terminal, create the database tables:
make migrateGood for development, because Go recompiles without rebuilding an image.
make db # start PostgreSQL
make migrate # create the database tables
make run # start the APIDo not use both options at the same time: they use the same port (8080).
The API listens on http://localhost:8080:
curl http://localhost:8080/healthz
curl http://localhost:8080/readyzBoth should answer 200 with {"status":"ok"}.
Run make help to see all commands.
make test # all backend tests
make test-race # same, with the race detector (needs a 64-bit C compiler)backend/ Go code (API, Dockerfile, database migrations)
compose.yaml PostgreSQL and the API for local development
Makefile shortcuts for the common commands
.github/ GitHub Actions workflows (CI)
MIT