Frequently asked questions
Quick answers about what bnrouter is and how it works.
What is bnrouter?
bnrouter is an AI API gateway. It aggregates multiple large-model providers behind a single OpenAI-compatible endpoint, so your application talks to one API instead of many.
How is it different from calling providers directly?
Instead of managing separate accounts, SDKs, and billing for every provider, you get one key, one endpoint, unified quotas, and one usage dashboard — while models remain interchangeable.
Is it free to use?
The gateway itself has no setup fee. You only pay for the model usage you consume, either via balance top-ups or quota-based subscription plans.
Is my data safe?
Yes. Requests are forwarded to the model providers over encrypted channels. We do not store your prompts or completions; only usage metadata is kept for billing and analytics.
Which models are available?
The catalog covers major chat, embedding, image, and video model families including Claude, GPT, Gemini, Kimi, DeepSeek, Qwen, Grok, Llama, Flux, DALL·E, Stable Diffusion, and more. Availability can vary by upstream channel, so check the model plaza in the console for the live list.
Do you support multiple languages?
Yes. The console and this site are available in English, Simplified Chinese, and Bahasa Melayu.
Does it work with OpenAI-compatible clients?
Yes. The endpoint implements the OpenAI chat and embeddings protocol, so the OpenAI SDK, LangChain, and other OpenAI-compatible clients work by simply pointing the base URL at bnrouter and using your key.
Can I start without a long-term commitment?
Yes. There is no setup fee and no minimum spend. Create an account, top up a small balance, and pay only for the usage you consume.