Your data
Your data stays in your account
Torque Lift is built so that what it produces belongs to you: the code in your repo, the events in your bucket, in formats any team can pick up. Here is what it can and can't touch.
Where events live
As Apache Iceberg tables in a bucket in your own cloud account. That copy is the system of record. Torque Lift's query layer reads from it and is never the only copy.What the agent can access
Your code, through a GitHub App on the repositories you choose. It can read code and open pull requests. It can't merge or approve them, push to your main branch, or touch workflows, secrets or settings.What the agent can't access
Your data. The agent that edits code runs in a sandbox with no data credentials and no GitHub token.Personal data
A blocklist of personal fields (emails, names, phone numbers, free text) is enforced in what the agent instruments and again in what the pipeline lands. Fields Torque Lift doesn't recognise are dropped or quarantined, never passed through.Your checks run first
Your own typecheck, tests and build run on every change in the sandbox before the pull request opens.Consent
The tracking Torque Lift writes sends nothing from a browser until your app grants analytics consent.When you leave
You keep the tables, the dbt project, the tracking code and the flags. Handing the stack to your own data team is a permissions change, not a migration.Certifications
Not yet. We're pre-launch, and we'd rather show you the controls than a badge we haven't earned. Ask for our security baseline when you join the waitlist.