Configure OpenCode with Earbowe API

OpenCode supports custom providers through opencode.json. You can point it at any OpenAI-compatible gateway — including Earbowe.

This guide follows the same pattern used by mature gateway docs (custom baseURL + @ai-sdk/openai-compatible).

1. Get a key

Create an API key in the Earbowe Console.

2. Choose base URL

Always include /v1.

3. Verify models first

curl -s https://api.earbowe.com/v1/models \
  -H "Authorization: Bearer sk-YOUR-KEY" | head -c 500

Copy exact model id values from the response. Do not invent marketing names.

4. Add a custom provider in OpenCode

Create or edit opencode.json (project or user config, depending on how you run OpenCode):

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "earbowe": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Earbowe",
      "options": {
        "baseURL": "https://api.earbowe.com/v1",
        "apiKey": "{env:EARBOWE_API_KEY}"
      },
      "models": {
        "gpt-4o-mini": {
          "name": "GPT-4o mini"
        },
        "deepseek-v4-pro": {
          "name": "DeepSeek V4 Pro"
        }
      }
    }
  },
  "model": "earbowe/gpt-4o-mini"
}

Then export the key:

export EARBOWE_API_KEY="sk-YOUR-KEY"

Notes:

• npm: @ai-sdk/openai-compatible is the usual adapter for OpenAI-compatible gateways.
• Model keys under models should match IDs returned by /v1/models.
• Default model format is often provider/model → earbowe/gpt-4o-mini.

5. China-direct variant

Only change baseURL:

"options": {
  "baseURL": "https://api.3861343.xyz/v1",
  "apiKey": "{env:EARBOWE_API_KEY}"
}

6. Common failures

Related

• Quickstart
• Codex setup
• Cline setup
• Troubleshooting