[code]smith
ProofPricing
You pay for what [code]smith uses. No seats.
Sandbox time only counts while [code]smith is working, so a minute on an 8 vCPU sandbox costs $0.016 and waiting on you is free. The per-million fee covers every token the model reads or writes, cached ones included.
Autofix runs are billed the same way. For Proof, see Proof pricing.
Keeping spend in check
- A session pauses the first time it reaches $200. Press Continue to keep going, or turn this off in Settings.
- Admins can set a monthly [code]smith spend limit in Settings.
- Long sessions re-read their whole context every turn. Start a new session for new work.
Billing
[code]smith bills incrementally for usage. The first time your combined monthly usage including runners crosses $100, you will be invoiced. If a charge fails, access to [code]smith may be restricted. CI runners will continue to run and accrue charges. You can restore access by paying your invoice from Usage and Billing.Data retention and sharing
We use a mix of model providers, set to zero data retention wherever the provider allows it. A few models, including the latest Anthropic ones, require the provider to keep data. Those stay off until an admin opts in, from [code]smith onboarding or the Settings page. Your prompts and code are only used to do the task you asked for. We don’t train on your code. To make cross-repo changes, [code]smith builds an index of the codebases you store with us. You can turn the index off in Settings. For the full privacy terms, see the privacy policy.Connector permissions
When you connect MCP tools, each server has an Availability scope in [code]smith:My Session (Personal)
Only your [code]smith sessions can use the connector. Credentials are stored for your user account and are not shared with teammates. Note that other users may still view output and sessions that use this server.
Team
Every org member’s [code]smith sessions can use the server. The org stores one shared OAuth or API credential.