Trust & data

Sangam-Git is the trust layer for AI-written code. This page states plainly how we handle your code and data. It is a summary of our practices, not a contract; formal Terms and a DPDP-compliant Privacy Policy accompany it.

Your code is yours

We host your repositories and never repurpose them. Data-residency options, including keeping storage and AI inference within a specific region, are available for teams that require them.

We never train on your code

Your source code, diffs, issues and pull requests are never used to train any AI model — ours or a provider's.

AI features are opt-out and transparent

AI features (pull-request summaries, review, config explanations) run only when you ask, on the specific diff involved. Each run shows the model used and its cost. You can turn AI off for your whole account in Account settings; provenance, risk flags and secret scanning keep working with no model calls.

Where AI runs

When AI is on, diffs are processed by our configured model provider under a no-training agreement. For customers who require it, an in-India inference option keeps model processing within the country — contact us.

Data retention & security logs

We retain operational and security logs for 180 days within India, in line with CERT-In directions. Deleting a repository removes its data from active systems.

Turning AI off

Account owners control AI per account in Account settings. When off, no code is sent to any model.