One familiar API
Use an OpenAI-compatible client or a simple HTTP request. Your application keeps its familiar structure.
Finding your route…
Chat, code, reason, create.
Bring your next idea to life with one powerful API.
v1Your first project starts with a single key.
Personal keys. Automatic routing. Built-in limits.
Requests for your next idea
A steady pace for building
Your own connection to every model
Three small steps. A whole new toolkit.
Create an account for your keys and usage.
Name it after your project. Save the secret when it appears.
Connect your app, choose a model, and send your first message.
Built for the things you want to make.
Use an OpenAI-compatible client or a simple HTTP request. Your application keeps its familiar structure.
Try your key in the playground. Carry the conversation forward before connecting your application.
Follow successful requests, tokens, and remaining allowances in your personal workspace.
Name, limit, and revoke your key. Your API secret is shown once, and your workspace stays private.
At 00:00 UTC. New accounts receive 100 successful requests per day and a limit of 10 attempts per minute.
Yes. Revoke your active key to free its slot, then create a replacement. Applications using the old key must be updated.
Start with smart, fast, coder, or reasoner, or copy a specific model ID from Model Status. Availability reflects the model catalog at the time of the check.