Run core-app locally - Dunsky Models & Tools

Use Docker Compose to start the core-app frontend, backend, local PostgreSQL database, and Models API together. The stack still needs the configured external services; it is not a fully offline demo.

Full stack

You need repository access, Docker Compose, the environment values required by azure.env.template, and permission to pull the Models API image from the private registry. From the repository root, copy azure.env.template to azure.env and fill it with approved development values. Never commit this file.

docker compose up --build

Open http://localhost:3000. The backend is also exposed on port 8081; PostgreSQL uses 5432 and the Models API uses 8080 by default. CORE_APP_PORT and CORE_APP_API_PORT override the app ports. Compose waits for PostgreSQL health, then runs tenancy and measures migrations and development seeding when the app starts. RUN_DB_MIGRATIONS_ON_START=true is a local-only convenience.

Run packages separately

The backend requires Python 3.15 or newer:

cd src/api
pip install -e '.[dev]'
runserver

The frontend pins pnpm in src/platform/package.json:

cd src/platform
pnpm install
pnpm start

Separate backend execution needs a reachable database and the configured Databricks services for operations that use them.

​Full stack

​Run packages separately

Full stack

Run packages separately