GitHub Copilot with Nbility
VS Code’s language model management supports Custom Endpoint models for chat and agents. This does not redirect all Copilot features, inline suggestions, or cloud services to Nbility.
Add a Custom Endpoint
- Update VS Code and its extensions, then open the Chat model picker.
- Select Manage Language Models → Add Models → Custom Endpoint.
- Enter a dedicated API key through the setup flow and configure the model in the opened
chatLanguageModels.json. - Retain the generated
${input:...}key reference so credentials stay in VS Code’s secret storage.
For Chat Completions, retain the wizard’s other required properties and fill in:
| Field | Value |
|---|---|
vendor | customendpoint |
apiType | chat-completions |
Model id | Exact ID from Models |
Model url | https://api.nbility.ai/v1/chat/completions |
The url is a complete endpoint. Selecting responses or messages also requires the matching full path and a compatible model.
Use and verify
Make the model visible in the picker, select it in Chat, and send a short message. Then try a small Agent task and check the actual requests in Usage logs.
Version and capability limits
- Use the current setup flow instead of the legacy
github.copilot.chat.customOAIModelssetting. - If Custom Endpoint is absent, check client version, organization policy, and feature availability.
- Set tool, vision, and context capabilities to match the real model. BYOK chat does not automatically provide inline suggestions or embeddings.
- In Responses mode, align conversation storage with the gateway; do not assume every compatible service stores
previous_response_id.
References
Adapted and rewritten for Nbility from aicentos’s GitHub Copilot guide, licensed under CC BY 4.0. Its account, plan, and proprietary service instructions do not apply here. See the client’s official documentation.