Integrate GitHub Copilot CLI

Use Copilot CLI BYOK environment variables with the Nusabase OpenAI-compatible endpoint.

GitHub Copilot CLI supports BYOK (Bring Your Own Key) for OpenAI Chat Completions-compatible endpoints. The Nusabase OpenAI-compatible Base URL is https://api.nusabasedigital.com/v1.

When To Use This

• You want to use Copilot CLI in the terminal.
• You want Copilot CLI model calls to go through Nusabase.
• Your selected model supports tool calling and streaming.

Copilot CLI BYOK requires the model to support tool calling/function calling and streaming. If the selected model or channel does not support these capabilities, Copilot CLI will return an error. Confirm available models with /v1/models, then choose according to model capability.

Prerequisites

• GitHub Copilot CLI is installed according to GitHub’s official docs.
• You have a Nusabase API key: sk-....
• Your machine can reach https://api.nusabasedigital.com.
• The example model qwen3.5-flash appears in the current pricing docs. Use /v1/models as the source of truth.

Check the CLI:

copilot --version

Configure Environment Variables

Linux / macOS / WSL:

export COPILOT_PROVIDER_TYPE=openai
export COPILOT_PROVIDER_BASE_URL=https://api.nusabasedigital.com/v1
export COPILOT_PROVIDER_API_KEY=sk-your_api_key
export COPILOT_MODEL=qwen3.5-flash

Windows PowerShell:

$env:COPILOT_PROVIDER_TYPE="openai"
$env:COPILOT_PROVIDER_BASE_URL="https://api.nusabasedigital.com/v1"
$env:COPILOT_PROVIDER_API_KEY="sk-your_api_key"
$env:COPILOT_MODEL="qwen3.5-flash"

Environment Variable	Description
COPILOT_PROVIDER_TYPE	Use openai for OpenAI Chat Completions-compatible calls.
COPILOT_PROVIDER_BASE_URL	Use https://api.nusabasedigital.com/v1.
COPILOT_PROVIDER_API_KEY	Nusabase API key.
COPILOT_MODEL	Default model name; must be enabled for your account.

If the model is not in Copilot CLI’s built-in catalog, set token limits explicitly:

export COPILOT_PROVIDER_MAX_PROMPT_TOKENS=128000
export COPILOT_PROVIDER_MAX_OUTPUT_TOKENS=8192

Verify API First

curl https://api.nusabasedigital.com/v1/chat/completions \
  -H "Authorization: Bearer sk-your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.5-flash",
    "messages": [{"role": "user", "content": "hello"}],
    "stream": true
  }'

Start Copilot CLI

copilot

Then send:

Introduce this project in one sentence.

Troubleshooting

Problem	Likely Cause	Fix
401	API key is wrong or not visible in the current shell	Set COPILOT_PROVIDER_API_KEY again in the same terminal.
404 or model unavailable	COPILOT_MODEL is not enabled for your account	Call /v1/models and copy a real model ID.
400	Provider type or Base URL does not match the CLI request format	Use openai and https://api.nusabasedigital.com/v1.
Tool calls fail	The selected model or channel does not support tool calling/streaming	Switch to a model that supports tool calling and streaming.
CLI still calls GitHub-hosted models	Environment variables were not inherited	Set them in the same terminal that starts copilot.

For VS Code workflows, see GitHub Copilot.

References

• GitHub Copilot CLI BYOK
• GitHub Copilot CLI product page
• First Request Example
• Models and Pricing
• Error Codes