Providers
TypeSafe
Call Jev directly at TypeSafe, the company behind the model.
The direct route: jev-guard sends each batch to TypeSafe’s API, and you pay TypeSafe for what you use.
Set it up
- Create an API key in the TypeSafe console.
- Give it to jev-guard, preferably as an environment variable:
export TYPESAFE_API_KEY=ts_...Or, if your host can’t set environment variables:
api:
provider: typesafe
typesafe:
key: "ts_..."- Run
/jevguard reload.
Settings
| Key | Default |
|---|---|
api.typesafe.base-url |
https://api.typesafe.ai |
api.typesafe.model |
jev-latest |
A direct check takes about 200 ms. For the model itself, see TypeSafe’s docs.