- Cut of your price
- 0%
- API formats
- 5
- Independent provider
- 1
- Statements
- Monthly
Apply
Company details and up to ten endpoints. A person reviews every application.
List models
Pick models from the catalog and set your price. A new model needs our approval once.
Pass the check
We send real requests before you go live. A failed recheck pauses you until you pass.
Get paid
A statement every month. Send us your invoice for it and we pay it.
What you need
- An HTTPS endpointPublic, so our gateway can reach it.
- One of five APIsOpenAI Chat Completions or Responses, Anthropic Messages, Gemini or Bedrock.
- An API keySealed the moment you save it.
- A data policyEU or not, training on prompts, days kept.
- Company detailsSo we can review you and pay you.
Questions
Who can apply?
Any company that serves models on a public HTTPS endpoint over one of the five APIs.
How do callers pick a provider?
Each request goes to the best healthy listing, ranked by price and speed. Callers can also ask for the cheapest first, or keep to providers in the EU or that keep nothing.
What if my endpoint goes down?
The request moves to the next provider, so the caller still gets an answer. Two failures in a row take you out of rotation for 30 seconds, and a failed check pauses your listing until it passes.
What does llm.buzz charge me?
Nothing. You earn your listed price on every token you serve. We charge callers a fee when they buy credit.
How am I paid?
When a month closes you get its statement. You upload your invoice for it, and we pay it.