OMENS / BUILD & UNDERSTAND

Run Omens locally

Use the same observatory on Windows, macOS or Linux.

Install and start

The project requires Node.js 22.13 or later, Git and pnpm 11.25.0, the version pinned in the repository.

git clone https://github.com/Wdustin1/omens.git
cd omens
pnpm install --frozen-lockfile
pnpm local

Open http://127.0.0.1:8787 in your browser. The launcher builds on first use and reuses production assets on later starts. The homepage, docs and app are available locally.

After source changes

Rebuild the local application after updating its source:

pnpm local --rebuild

For development with hot reload, use pnpm dev. Local clients default to Low power mode. External feeds, map tiles and OpenAI requests still need internet access.

Optional service scaffold

The repository includes a separate service protocol scaffold. It is not a proprietary forecasting engine. Start it in another terminal:

pnpm service

In Omens AI, select Omens-compatible service / device and connect to http://127.0.0.1:8788 from the local application. The expected state is Connected · prediction engine unavailable. Scenario requests return HTTP 501 until an engine is implemented.

The OpenAI option is separate and can be used without this scaffold. Connecting the hosted app to a device requires an explicitly configured authenticated HTTPS service with the correct CORS origin.

Work with the source

The GitHub repository ↗ contains application code, adapters, validation scripts and technical protocol notes. Never commit provider keys, service tokens or wallet credentials.

The hosted website uses a separate publication process. A GitHub push alone does not deploy it.