> For a complete documentation index, fetch https://docs.voximplant.ai/llms.txt

# Overview

> For the complete documentation index, see [llms.txt](/llms.txt).

[Jev](https://docs.typesafe.ai/) is TypeSafe's decision engine for structured evaluations over text and images. **System One** runs a single decision: you provide a `state` (for example, ticket text) and a set of **questions** (yes/no, choice, score, and related types), and the model returns typed **answers** you can use in IVR logic, routing, or agent workflows.

In your VoxEngine [scenario](/platform/voxengine/scenarios), require **Modules.Jev** and use a [Jev.SystemOneClient](https://voximplant.com/docs/references/voxengine/jev/systemoneclient). One client keeps a WebSocket session open so you can send several [systemOne](https://voximplant.com/docs/references/voxengine/jev/systemoneclient#systemone) requests without reconnecting. This path is **not** a realtime voice pipeline: you do not send call audio to Jev; you send JSON request bodies and read JSON results.

## See also

* [System One client](/voice-ai-orchestration/jev/system-one-client) — scenario setup and example
* [TypeSafe System One concepts](https://docs.typesafe.ai/concepts/system-one)
* [Voice AI getting started](/getting-started/voice-ai-on-voximplant/overview)