OpenClaw + OpenBulu configuration
For versions that support a custom OpenAI-compatible provider, use the Chat Completions endpoint and an exact model ID.
Protocol: OpenAI-compatible. Base URL: https://openbulu.com/v1. Endpoint: POST /v1/chat/completions.
Configuration reference
{
models: {
providers: {
openbulu: {
baseUrl: "https://openbulu.com/v1",
apiKey: "${OPENBULU_API_KEY}",
api: "openai-completions",
models: [{ id: "<MODEL_ID>", name: "<MODEL_ID>" }]
}
}
},
agents: { defaults: { model: { primary: "openbulu/<MODEL_ID>" } } }
}Configuration and verification steps
- Add an openbulu entry under models.providers in the OpenClaw configuration. Set baseUrl to https://openbulu.com/v1, api to openai-completions, apiKey from OPENBULU_API_KEY, and models to objects with exact model IDs.
- Set agents.defaults.model.primary to openbulu/<MODEL_ID> and confirm that the selected model uses the OpenBulu provider; check existing agent model overrides before testing.
- Send a small agent request and match its time, model, and /v1/chat/completions route in OpenBulu usage logs. Test tools and fallbacks separately.
Use your own API key and an available model ID. Confirm the selected provider and model in the client, run a small request, and match its time, route, and model in your OpenBulu usage log. Advanced tool actions may need separate testing.