SECURITY

How Sabí handles your code and data.

Plain answers for the person who has to say yes.

Colleagues smiling around a table with laptops
OUR PROMISE

Your code visits. Your docs stay.

Each run analyses your code in a secured sandbox that's deleted when the run ends.

Every run gets its own ephemeral, secured sandbox: read only, with no network access. Your full repository is never sent to the Sabí engine. When the run ends, the sandbox is deleted, whether the run worked or not. We keep the documentation and the short quotes it cites.

WHY TEAMS SAY YES

What you can count on.

Reads, never writes

Sabí can look at your code but never change it.

What we keep

The documentation and the short code excerpts it cites. Not your repository.

Only what's needed

For each page, the Sabí engine works with the parts of your code that matter. Your full repository is never sent to the Sabí engine.

Every page shows its state

Backed by evidence, checked in the product, edited by a person, or failed.

Product checks stay in test

The beta product check only visits the test addresses you register. Never your live product.

Honest about GitHub

Sabí's own code is stored on GitHub. Yours never is.

Not used for training

Sabí doesn't train AI models on your code.

Anchored to your code

Every page is written from the code itself, and shows the change it came from.

WHERE WE ARE

Certification, honestly.

Data processing agreementAvailable now
Subprocessor listPublished
Penetration testPlanned before our first pilot
ISO 27001On our roadmap
ISO/IEC 42001 (AI management)On our roadmap
Single sign-onOn request
SECURITY PACK

Have a security review to run?

Send us your questionnaire, or ask for the pack: architecture summary, completed questionnaire and our data processing agreement.

Book a demo
Security deep dive, for technical reviewers
A Sabí teammate