LM Studio Setup (Custom / Local)

Use this guide when you want Genie to talk to a local model server or any OpenAI-compatible gateway instead of a hosted provider account.

What this setup assumes

Genie does not run the model itself. You need a server that already exposes an OpenAI-compatible API.

Typical options:

Open Settings and look at the provider settings for:

You can point both roles at the same endpoint or split them across different models.

When you choose Custom / Local, Genie exposes these fields for each role:

Example endpoint shape:

http://127.0.0.1:1234/v1

Set Coding Provider to Custom / Local.
Enter the model ID your local server expects.
Enter the endpoint base URL.
Add the API key only if your server requires one.
Click Test Connection before saving.
Repeat for Logistics if you want ASK, QUERY, or DATA to use the same local stack.

Check these first:

Activation and provider credentials are separate.

Fix activation in Licensing and Activation only when the UI is locked.
Fix provider configuration in AI Providers or Settings Guide when the UI is available but model calls fail.

Check Coding and Logistics separately. Each role can have its own endpoint, model, and key.