# Can I self-host Jev?

No. TypeSafe serves Jev only through its own hosted API and through OpenRouter's hosted endpoint for it; neither its docs nor its terms of service mention a self-hosted build, downloadable weights, or an on-premises option. TypeSafe's terms define the product itself as "the TypeSafe-hosted web interface" and "the TypeSafe-hosted application programming interface," and every version of the model is served from the same endpoint, `POST /v1/systemone`, selected by name in the request rather than installed anywhere. That matches how Tessary itself reaches Jev to run its frustration classifier: through a TypeSafe key or an OpenRouter key, never a model file. [You still need to supply one of those keys yourself](/answers/tessary-classifiers/do-i-need-my-own-llm-provider-key-for-tessary-to-grade-traces), because there's nothing in Jev's case to run locally instead.

---

Sources:
- TypeSafe Master Customer Agreement: https://typesafe.ai/legal/mca (fetched 2026-10-01)
- TypeSafe docs, Models: https://docs.typesafe.ai/models (fetched 2026-10-01)

Source: https://tessary.ai/answers/jev/can-i-self-host-jev
More on Jev: https://tessary.ai/answers/jev
From Tessary, agent reliability for AI agents in production: https://tessary.ai