Shimpilot is an independently operated software service. In this policy, "Shimpilot", "we", and "us" refer to the people operating the service. Shimpilot has not yet been incorporated as a registered company. For questions about this policy, contact privacy@shimpilot.com.

This policy covers the Shimpilot web application, its API, its background scanning and migration workers, and its GitHub App integration. It does not cover GitHub, your API providers, or any third-party service you separately use.

Account information. Your email address and authentication identifiers, handled through our authentication provider.

Organization information. The organization record associated with your account, its subscription status, and its alert configuration (Slack webhook URL and alert email address, where you provide them).

GitHub installation information. The installation identifier for the Shimpilot GitHub App, and the repositories you explicitly grant it access to.

Repository metadata. Repository name, default branch, visibility, scan timestamps, resolved dependency evidence, and coverage state.

Code analysis results. File paths, line numbers, function names, and the API call symbols detected in your repository, together with the detected SDK version. These are derived facts about your code, and they are stored.

Migration records. Generated patches, validation results, and pull request URLs and numbers.

API provider metadata. Publicly published changelogs, specifications, and release notes retrieved from API providers. This contains no customer data.

Diagnostic information. Server-side logs recording job identifiers, repository identifiers, timing, and error classifications.

Website usage information. Google Analytics may process page views, referral sources, approximate location, and broad device or browser information when you use Shimpilot's website.

To analyse a repository, Shimpilot performs a shallow clone into a temporary working directory, parses it, and deletes that directory when the job finishes — including when it fails. We do not operate a long-term store of your repository contents.

We want to be precise rather than reassuring here, because two categories of code-derived data are retained beyond the scan:

• Code-location evidence. Findings store the affected file path, line, column, symbol, reviewed rule identifier, and version evidence. The scanner does not persist full source files as finding snippets.
• Generated patches. A migration record stores the generated diff, which by its nature contains fragments of your code before and after the change.

So: findings retain structured location evidence rather than source excerpts. Generated migration patches may still contain fragments of source code before and after the proposed change. We do not retain whole files or a permanent copy of the repository.

To detect upstream API changes, identify affected code in your repositories, assess migration risk, generate and validate migration patches, open draft pull requests, send the alerts you configure, operate and secure the service, and handle billing where applicable. We do not sell personal information or train our own models on your source code. Third-party AI providers process the data sent to them under their own terms and policies, which you should review before enabling migration generation.

Shimpilot sends relevant findings and repository metadata to Groq for summaries and migration planning. For code generation and repair, the configured AI endpoint may receive affected source files and, when applicable, typecheck errors, test output, and CI diagnostics. We do not intentionally include platform credentials. Automated secret screening cannot guarantee that customer source is free of embedded secrets; do not commit credentials to connected repositories.

AI output is treated as a proposal, never as an authority. A generated patch reaches a pull request only after deterministic validation. Manual mode leaves the merge decision to you; Automatic mode can merge only after the non-AI safety gates authorize it. Shimpilot does not deploy your application.

Data is stored in our managed Postgres database and job queue. Traffic between you and Shimpilot, and between Shimpilot and its providers, is encrypted in transit. Repository access uses short-lived GitHub App installation tokens rather than long-lived personal access tokens. Platform credentials are held as environment configuration. Newly saved Slack webhook URLs are encrypted at the application layer; older URLs remain readable until migrated. We do not intentionally include secrets in logs, pull requests, or AI prompts.

No service can promise absolute security, and we do not. See our Security page for a factual description of the controls that exist today.

Shimpilot relies on third-party providers for hosting, database and authentication, background job processing, AI patch generation, email delivery, and billing. Each receives only what its function requires. Our subprocessors page lists them.

Temporary repository clones are deleted at the end of each job. Stored findings, dependency records, impact analyses, migration records, and diagnostic logs do not currently have a published fixed retention period. Retention depends on the record type and the storage provider. Contact privacy@shimpilot.com to ask about retained data or request deletion.

Depending on where you live, you may have rights to access, correct, export, or delete your personal information, and to object to or restrict certain processing. Contact us at privacy@shimpilot.com and we will respond within the period required by applicable law.

You can revoke Shimpilot's access at any time by uninstalling the GitHub App, which immediately ends our ability to read your repositories. To delete your account and its associated records, contact privacy@shimpilot.com.

Removing a repository from the GitHub App installation stops future scans but does not by itself delete previously stored findings — this is deliberate, so re-adding a repository does not lose its history. Ask us if you want those records erased.

Our providers may process data in countries other than yours. Where required, we rely on appropriate safeguards for those transfers. Specific processing locations are listed on the subprocessors page where we have confirmed them.

Shimpilot is a developer tool intended for use by adults in a professional capacity. It is not directed to children, and we do not knowingly collect personal information from anyone under 16. If you believe a child has provided us information, contact privacy@shimpilot.com.

We may update this policy as the product changes. Material changes will be reflected in the "Last updated" date at the top of this page, and where appropriate we will notify account holders directly.

For privacy questions, data requests, or questions about this policy, email privacy@shimpilot.com.