> ## Documentation Index
> Fetch the complete documentation index at: https://docs.warmcore-tech.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Audio and LLM integration

> Prototype with explicit boundaries

> **Developer Preview / Hardware Not Yet Verified:** X3 hardware and the deployed robot runtime have not been validated. Local tests do not establish device compatibility, physical behavior, or safe operation.

## Text model integration

The preview adapter accepts a compatible chat-completions endpoint through `LLM_CHAT_ENDPOINT` and a model through `LLM_MODEL`; use `LLM_API_KEY` only when required. Configure secrets through your environment or approved secret manager, never source code or shared logs. Keys require HTTPS except on loopback. Plain HTTP also exposes conversation text; use it only on loopback or an explicitly approved protected network.

From the authorized SDK checkout, the following requests a text response on the workstation:

```bash theme={null}
python3 examples/llm_robot_reply.py "Hello"
```

This requires a configured compatible provider. Live provider compatibility remains unverified. Sending output to the robot can affect playback and requires separate operator approval.

## Audio scope

Experimental PCM utilities are available, but playback requires a compatible authorized interface package and device validation. The preview does not replace the complete microphone-to-model voice pipeline or guarantee lip synchronization. Protect recordings and conversation content.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.