Warp with Nbility
Warp custom inference endpoints depend on release and account availability. This page applies when AI settings allow a custom model or inference destination.
Configuration entry point
Open AI model settings, add a Custom model / Inference endpoint, and select OpenAI compatibility. Save the endpoint, dedicated key, and model ID, then choose it in a new Agent session.
Connection details
| Field | Value |
|---|---|
| Type | OpenAI Compatible / Chat Completions |
| Base URL | https://api.nbility.ai/v1 |
| API key | Dedicated token from Token management |
| Model ID | Exact ID from Models |
This table applies to a base URL field that appends /chat/completions. Follow the field documentation when a complete endpoint is requested; see Connection basics.
Verification and limits
If Warp routes requests through a remote service, that service must reach the endpoint; localhost on your computer will not work. BYOK does not establish pricing or account requirements for Warp’s other features. Test text before terminal tools.
Send a short message in a new conversation and check the request and model in Usage logs. With no log, inspect the active provider and network first. With a failed request log, use the error guide.