driftfix Privacy

What happens to your code

Driftfix reads your repository to find code that an API change breaks, and opens a pull request that fixes it. This page says exactly what it reads, where that goes, what it keeps, and what API providers can see.

We read your code only to find and fix usages of a changed API.
Only the affected files are sent to our AI model provider, never the whole repository.
Your code runs only in a throwaway container with no secrets, which is deleted after each run.
We don't keep your source code. Clones are deleted after every job.
API providers see aggregate counts only, never your code, and never your repo's name unless you opt in.
We never merge, never push to your default branch, and never ask for access to your secrets or workflows.

Permissions we ask GitHub for

Contents: read & write, to read your code and push a branch named driftfix/… for the pull request.
Pull requests: read & write, to open the pull request and notice when it's merged or closed.
Metadata: read, which GitHub requires for every app.

Nothing else. We don't request administration, workflow, secrets, issues, or organization permissions. GitHub issues our access tokens per installation, and they expire within an hour.

What happens when an API you use changes

We clone your repository into a temporary folder on our server and search it for code that uses the changed API.
If nothing uses it, we stop there. Nothing is sent anywhere.
If something does, we send the affected files (and files that import the provider's SDK), along with the provider's published changelog and migration guide, to our AI model provider, Anthropic, to write the fix. We use Anthropic's API, which doesn't use API inputs to train models by default.
We check the fix mechanically. It may only touch the affected files, and never CI configuration, workflow files, environment files, or install scripts. It may not add new network calls. The only dependency change Driftfix makes is upgrading the provider's own SDK when the fix needs it.
We run your tests and type check, before and after the fix, in a throwaway container: no GitHub token, no API keys, none of your environment variables, and no network access while your code runs. The container is deleted afterwards.
We open a pull request on a new branch. You review it and decide. We never merge.
We delete the clone.

What we keep

Your GitHub account or organization name and installation ID.
For each repository: its name, default branch, main language, which tracked API providers it uses, your Driftfix settings, and whether you've chosen to share its status with providers.
For each run: its status, the pull request link, whether tests passed, and a step log with timings, token counts, and short error messages, which can include file paths.

We don't store your source code, diffs, or test output on our side. Diffs and test output live in the pull request on GitHub, in your repository.

What API providers can see

API companies pay Driftfix to get their customers migrated. You don't, and your privacy doesn't depend on whether they do.

By default, a provider sees aggregate counts per change: how many installed repositories were affected, how many pull requests were opened and merged, and why others are blocked. Counts below five are hidden so they can't point to a single customer.
A provider sees your repository's name and its pull request status only if you opt in. Opting in is off by default and you can turn it off at any time.
You can also leave a repository out of provider totals entirely, in your Driftfix settings. It then isn't counted anywhere, and can't be named.
Providers never see your code, diffs, file paths, or test output, whatever your settings.

Turning it off and deleting your data

To skip one provider, add {"ignore": ["openai"]} to a .driftfix.json file in your repository.
To pause Driftfix for a repository, turn off “Open pull requests” in your Driftfix settings, or remove the repository from the app's access in your GitHub settings.
Uninstalling the app deletes your installation, its repositories, and their run history from our database.

Cookies

When you sign in, we set one cookie that keeps you signed in for up to 8 hours. It's encrypted and can't be read by scripts on the page.
The demo stores your demo settings and tour progress in your own browser only.
We don't use advertising or analytics cookies, and no third-party scripts run on our pages.

If you contact us

The contact form stores your email address, your message, and whether you're a developer or an API company, so we can reply. We don't store your IP address with it. Ask us and we'll delete it.

Questions

Use the contact form on our home page and we'll reply. If we change how we handle your code, we'll update this page and the date below before the change takes effect.

Last updated October 6, 2026. See also our Terms of Service.