Pay for what you use. Nothing else.
No subscription and no plan to pick. Each project has its own balance, and what the assistant uses is deducted from it.
Token prices
Per million tokens. A token is about three quarters of a word.
Input
What the assistant reads: your request and the code it looks at.
$0.12 / 1M tokens
Output
What the assistant writes: its answers and the code it changes.
$0.60 / 1M tokens
Cached input
Code it already read earlier in the session.
$0.012 / 1M tokens
Cache write
Input kept for a few minutes, to be read again for less.
$0.15 / 1M tokens
Cache write (1 hour)
Input kept for an hour.
$0.24 / 1M tokens
Questions you are about to ask
- Is there a free plan?
- No, but every new project starts with $1.00 of credit, enough to try your first requests before adding a card.
- What do I pay for?
- The tokens the assistant uses to read your code and write changes, at the prices listed on the pricing page. There is no subscription and no commitment.
- Who pays: me or my collaborators?
- The project. Each project has its own balance, topped up by card, and the usage of every member is deducted from it.
- What happens when the balance runs out?
- The assistant stops until someone tops the project up. Turn on automatic top-up and the saved card recharges it when it gets low.
- Can I cap spending?
- Yes. Set a monthly limit for the project and one for each member: the assistant stops when it is reached.
Try it on your own app.
Create an account, connect a repository, and spend your first $1.00 on real changes.