Skip to content
Typillar Typillar Typillar

Models & API keys

Typillar generates code with a model that runs on your side. You choose how: use Cloudflare Workers AI, or bring your own provider key. Either way, no inference runs on Typillar’s servers — see Zero inference on our servers.

Option 1 — Cloudflare Workers AI

Section titled “Option 1 — Cloudflare Workers AI”

If you’ve connected Cloudflare, you can run inference on Workers AI with no extra setup. Tokens are generated on Cloudflare’s network and billed through your Cloudflare account. This is the simplest path and a good default.

Option 2 — Your own provider key

Section titled “Option 2 — Your own provider key”

To use a specific model or vendor, supply your own API key in the console. Typillar will route inference for your builds through that provider.

How your key is stored

Section titled “How your key is stored”
  • Provider keys are encrypted at rest (AES‑GCM) — they are never stored in plain text.
  • A key is used only to run inference for your builds.
  • You can rotate or remove a key at any time from the console; removing it stops further use immediately.

Switching later

Section titled “Switching later”

You can start on Workers AI and move to your own key (or the reverse) whenever you like. It changes only where the tokens are generated — the build loop, your approvals, and your deployments are unaffected.

Section titled “Related”