AI-Change Provenance
A tamper-evident record tying each code change to the model call, tool calls, and authorization behind it.
This capability is on our roadmap — not shipped today. Early access opens to a small group at a time; requesting a spot is free and carries no commitment.
Request early accessThe problem
Auditors and security reviews are starting to ask a question most teams cannot answer: which of these changes did an AI agent make, and who authorized it? Git history shows what changed, but not the model invocation, the tool calls, or the grant that allowed it. When something goes wrong, reconstructing that chain after the fact is guesswork — and guesswork is not evidence.
What you'd get
- Every AI-assisted change linked to the model call, tool calls, and authorization that produced it.
- Tamper-evident records — evidence that stands up in an audit, not a log that could have been edited.
- Answers on demand: which agent, authorized to do what, delegated by whom, for any change in your history.
- Local-first capture that works alongside the reviews and tools you already run.
Who it's for
- Security and compliance leads preparing for audits that now ask about AI-generated code.
- Platform engineers who need post-incident forensics for agent actions.
- Teams that want to say yes to AI agents and still prove exactly what those agents did.
Get early access to AI-Change Provenance
A tamper-evident record tying each code change to the model call, tool calls, and authorization behind it — evidence you can hand an auditor.