Korvexor
Get started

OpenCode setup

Add a Korvexor provider in opencode.json using the OpenAI-compatible /v1 endpoint.

  1. Create OpenCode config

    Add opencode.json in your project or global config with a korvexor provider.

  2. Fill baseURL and apiKey

    Set options.baseURL with /v1, apiKey from the console, and model to korvexor/auto.

  3. Run OpenCode

    Start opencode and confirm model discovery and your first request work.

OpenCode provider config diagram

Example config (copy-ready):

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "korvexor": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Korvexor",
      "options": {
        "baseURL": "https://api.korvexor.com/v1",
        "apiKey": "sk-your-api-key"
      },
      "models": {
        "auto": {
          "name": "Korvexor Auto Router"
        }
      },
      "discoverModels": true
    }
  },
  "model": "korvexor/auto"
}

Notes

  • • baseURL must end with /v1 (OpenAI Compatible SDK convention).
  • • Claude Code uses the gateway root without /v1—the two rules are not interchangeable.

Next steps

Try cURL examples or the troubleshooting page to validate protocol and URLs.

cURL examples →Troubleshooting →