Command Code with Nbility

Command Code accepts custom BYOK providers. This example uses OpenAI Chat Completions; models requiring Responses need openai-responses instead.

Register the providerCopy link

Create a dedicated token and set it in the launch terminal:

export NBILITY_API_KEY="YOUR_API_KEY"

Merge a provider into ~/.commandcode/providers.json:

{
  "provider": {
    "nbility": {
      "name": "Nbility",
      "api": "openai-completions",
      "baseURL": "https://api.nbility.ai/v1",
      "apiKey": "$NBILITY_API_KEY",
      "models": { "YOUR_MODEL_ID": {} }
    }
  }
}

Replace YOUR_MODEL_ID with an exact ID from Models. Keep the environment reference in apiKey; do not commit a real key.

Select and verifyCopy link

Use /connect to select Nbility and /model to choose its model. Reopen the model selector after editing the provider file.

Send a short message and check Usage logs. Listing a model does not verify its tool support; coding tasks require a suitable model.

Protocol and addressCopy link

• openai-completions: https://api.nbility.ai/v1.
• openai-responses: the same /v1 base, with a Responses-compatible model.
• Native Anthropic: create a separate anthropic-messages provider using https://api.nbility.ai.

Match the protocol to the model’s exposed capabilities rather than changing the protocol merely to switch model names.

ReferencesCopy link

Adapted and rewritten for Nbility from aicentos’s Command Code guide, licensed under CC BY 4.0. Its account, plan, and proprietary service instructions do not apply here. See the client’s official documentation.