Genie
Documentation

Genie Docs

Two ways in: the products (Rollup, Review, Builder, and the platform pieces) and the API — an OpenAI-compatible REST endpoint on open-weights models running on local GPUs. Point your existing OpenAI SDK at https://api.genie.tech/v1 and go.

Products

Endpoints

Get a key

Sign up, enable inference on your org, and mint a key under /settings/api-keys. 375 signup credits free. Org admins can mint keys for team members and scope them to per-project budgets — see Budgets.

What you can build

  • Drop-in OpenAI replacement for chat / completions / tools.
  • Open-weights image + video gen for products you don't want sending pixels to closed APIs.
  • Async batch jobs (financial filings, contract summaries, overnight runs) with HMAC-signed completion webhooks.
  • Multi-tenant — split your org into Applications (e.g. fedoverwatch, sup.video) with separate caps and usage analytics.

Where your data goes

Your data — and your users' data — never reaches a third party. Every request is processed by Genie, a US company, on infrastructure we run; it is never sold and never used to train a model. That holds because Genie has no third party behind it: every model is open-weight and runs on our own GPUs, so there is no closed-weight API to fall back to under load. See the enterprise page for the data-handling contract.