Stable access to real models
An integration guide for professional developers. We do not compete on the lowest price; we focus on real models, stable routes, and transparent usage.
Make your first request in six steps
- Create an openbulu account. Register and open the console to review current trial benefits.
- Create an API key. Use separate keys for projects so quotas, permissions, and logs stay clear.
- Top up your wallet. Complete a top-up of at least $10. Wait for the order to show successful payment before calling the API.
- Choose a model available to your key. Query /v1/models with your key and copy an exact model ID. A model listed in pricing may still be restricted for that key.
- Send a test request. Use cURL, PowerShell, the OpenAI SDK, or a compatible client with your key and an allowed model ID.
- Check Usage Logs. Confirm the post-payment request, model, tokens, and charge in Usage Logs.
Available API protocols
These are the protocol routes documented for openbulu. Use the Base URL shown in the console and do not append the path twice.
- OpenAI Chat Completions. The broadest default route for ordinary chat, streaming, and common tool-calling scenarios. Start here when you are unsure.
- OpenAI Responses API. A route for the Responses API ecosystem, including clients built around input, output, and response events. Exact parameters still depend on the model.
- Anthropic Messages. The message-format route for the Anthropic SDK, Claude Code, and similar clients. Headers and model capabilities depend on the client and model configuration.
- Gemini-style route. For clients that send Gemini-style generateContent requests. Confirm the model name and supported capabilities in the Model Center.
Where to go next
See client integration examples for tool-specific base URLs and verification checks. Compare current model pricing and contact support if a request fails.