# TypeSafe

> Configure TypeSafe deployments for the Jev decision models.

URL: https://boelabs.com/en/docs/ai-gateway/providers-typesafe
Language: English

## Configure a deployment

Send this body to `POST /admin/deployments` with an operator credential. Replace the example
credentials. See [Creating deployments](https://boelabs.com/docs/ai-gateway/creating-deployments) for the shared configuration
fields.

```json
{
  "publicModel": "triage",
  "adapterKey": "typesafe",
  "upstreamModel": "jev-1.14.0",
  "credentials": { "apiKey": "..." }
}
```

## Behavior and limitations

- **Credentials**: `apiKey` is required. `baseUrl` and custom `headers` are optional. The default base
  URL is `https://api.typesafe.ai`.
- **Operation**: decisions only. The adapter sends `POST /v1/systemone` through the `systemone`
  transport, whichever public contract the request came in on.
- **Models**: `jev-1.14.0`, `jev-1.13.0` and the `jev-latest` alias, which TypeSafe moves to its newest
  stable release. Only Jev 1.14 answers `box` questions.
- **Images**: inline PNG or JPEG only, up to 16 MiB each; image URLs are not accepted. The gateway
  sends them as `assets` and adds `TypeSafe-Danger-Zone-Images: true` to requests with images or box
  questions.
- **Long context**: Jev reads 65,536 tokens per request by default. To raise it to 262,144, set
  `"headers": { "TypeSafe-Danger-Zone-Long-Context": "true" }` in the deployment's credentials.
- **Input**: Jev takes a single state, so conversations and tool items are flattened into one labeled
  input, and audio, video and documents are not accepted.
- **Pricing**: $0.042 per million input tokens; output is free.

## Next steps

- [Decisions](https://boelabs.com/docs/ai-gateway/decisions) — the three public contracts and how they translate.
- [Model catalog](https://boelabs.com/docs/ai-gateway/model-catalog#decisions) — the `decision.create` profile.
- [Routing](https://boelabs.com/docs/ai-gateway/routing) — retries, fallback, health and price strategy.