Solo
Free
One repository, bring your own everything.
- 1 repository
- Up to 3 active pull requests at a time
- Bring your own AWS Bedrock and GitHub App
- Per-PR cost caps and attribution
Introductory pricing
Priced for a dollar per pull request, not twenty.
You bring your own AWS Bedrock, at cost, with per-PR attribution — Flight provides the execution sandboxes and the safety net around them. These are our launch numbers; they will move as we learn from real fleets, never mid-billing-cycle without notice.
Every tier brings its own AWS Bedrock. Model spend is yours, at cost, itemized per PR.
Team
$249
per repository / month
For a team shipping AI-authored PRs every day.
- Unlimited active pull requests
- Every Solo feature
- Shared cost caps and budget alerts across the fleet
- Priority support
Enterprise
Custom
license Flight inside your org
Self-hosted, custom limits, dedicated support.
- Self-hosted deployment option
- Organization-wide cost governance
- Custom SLAs and onboarding
- Volume pricing across repositories
Cost caps are built in, not bolted on.
You set the maximum a single pull request may spend. Flight tracks spend against it in real time, in dollars, attributed to that exact PR in your AWS bill. No surprise invoices, cancel anytime.
Frequently asked
Who pays for the AI model spend (Bedrock)?
You do, at cost, through your own AWS account. Flight never marks it up — every dollar is visible per pull request in your AWS bill.
Who provides the execution sandboxes?
Flight does. Every labor and judgment step runs in an isolated sandbox we provision and pay for as part of your subscription.
What happens when a pull request hits its cost cap?
The PR freezes and Flight leaves a precise, actionable recommendation for a human. Spend never continues silently past the cap you set.