OpenBulu
Taking longer than expected? Reload this page.

OpenCode

Prefer an OpenAI-compatible provider and enter the openbulu Base URL, key, and model ID. Advanced parameters depend on the client version.

Protocol: OpenAI Responses (1.x). Base URL: https://openbulu.com/v1. Endpoint: POST /v1/responses.

Configuration example

OpenCode 1.18.35 (2.x uses a different schema)
PowerShell: $env:OPENBULU_API_KEY = "<OPENBULU_API_KEY>"
Bash/Zsh: export OPENBULU_API_KEY="<OPENBULU_API_KEY>"
Create opencode.json in a disposable project folder:
{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "openbulu": {
      "npm": "@ai-sdk/openai",
      "name": "OpenBulu",
      "options": {
        "baseURL": "https://openbulu.com/v1",
        "apiKey": "{env:OPENBULU_API_KEY}"
      },
      "models": { "<MODEL_ID>": { "name": "OpenBulu model" } }
    }
  }
}
Run: opencode run --model openbulu/<MODEL_ID> "Reply with exactly: OpenBulu connection successful"
Check the /v1/responses call in OpenBulu Usage Logs.

Examples use placeholders only. Keep your real API key in an environment variable or the client’s secure credential store.

Setup steps

  1. Use OpenCode 1.18.35 for this tested path. Set OPENBULU_API_KEY in the current shell, create opencode.json in a disposable project folder from the example, and replace <MODEL_ID> with an exact available ID.
  2. Run the example opencode run command and confirm the reply is exactly OpenBulu connection successful.
  3. In OpenBulu Usage Logs, match the request time, model ID, and /v1/responses route. Test file editing separately in a disposable folder and run its test yourself.

Verify setup

Create a key, choose a model available to your account, run a small request in the client, then check its model, result, and charge in Usage Logs. Do not share the full key.

OpenCode 1.18.35 on Windows completed an isolated calculator file-edit task; its independent Node test passed, with matching streaming /v1/responses logs for gpt-5.6-sol. This evidence does not verify another version or model.

Troubleshooting

All client guides · Model pricing