Windsurf with a custom OpenAI-compatible provider

Windsurf accepts a custom provider in its model settings. Add the base URL, the key and an exact catalog id, and Cascade runs on this catalog.

Base URL

https://uncensoredai.dev/v1

01
Open the model settings
Open Windsurf settings and find the model section. Choose Add custom model and select the OpenAI-compatible provider type.

Enter the endpoint and key

Paste the base URL, the access key and a catalog model id. Keep the /v1 suffix.

Custom provider

Provider:  OpenAI compatible
Base URL:  https://uncensoredai.dev/v1
API Key:   sk-ua-XXXX-XXXX-XXXX-XXXX
Model:     deepseek-v4.1-flash

03
Test in Cascade
Send one prompt in Cascade. If you see a 404, the model id is wrong; a 401 means the key was not saved.

Commands and configuration below are shown exactly as the tool expects them, so they stay in English.

Do I lose Windsurf's built-in models?+

No. Custom providers sit alongside them; you choose per conversation which model to use.

Is autocomplete covered?+

Custom providers cover chat and Cascade. Supercomplete and tab autocomplete remain on Windsurf's own infrastructure.