# Reference (/reference)

The guides explain _how_ and _why_. These pages are the exact listings to look
things up in. The config, provider and model tables are generated from the same
schema and catalog the CLI and engine use, so they match the code.

<Cards>
<Card title="config.cuelo" href="/reference/config">
Every field with its type, default and meaning, plus a complete example.
</Card>

<Card title="CLI" href="/reference/cli">
Every `cuelo` command and flag, the Python entry points, exit codes.
</Card>

<Card title="Providers" href="/reference/providers">
Speech, LLM and turn-taking providers: class, extra, credential, supported settings.
</Card>

<Card title="Models" href="/reference/models">
The turn-taking models Cuelo runs locally: license, size, CPU cost, status.
</Card>

<Card title="Python API" href="/reference/python">
Harness, Session, Pipeline, the slot interfaces, tools and serving.
</Card>

<Card title="Environment variables" href="/reference/environment">
Credentials and every `CUELO_*` setting.
</Card>
</Cards>