Where your documents go.
What SecondSight keeps, what it sends, and what it doesn’t claim.
What runs where
Our servers
The app, the database (PostgreSQL), and the document reading: OCR and layout models on our own GPU.
AWS S3, Mumbai (ap-south-1)
Your original files and their outputs.
Every outgoing connection
| To | When and what |
|---|---|
| AWS S3 (ap-south-1) | Your files and outputs, stored. |
| Google Gemini | Figure images, to write descriptions; title and cover pages, to read the title details. |
| Cartesia | The words of passages read aloud, to make the voice. Audio is kept on our server so the same words aren’t sent twice. |
| Anyone else | Nothing. |
Sign-in
Each person signs in with an email and password. Passwords are stored hashed (bcrypt). What each person can open is set by roles and module access.
On-premise
Everything runs on your server. Gemini and Cartesia are used only if you give it keys for them. Without them, figures aren’t described by AI, and Read-Aloud uses the browser’s voices.
What we do not claim yet
No SOC 2 or ISO 27001 certification. No single sign-on and no two-factor sign-in. If your review needs any of these, tell us.
Questions from your security review?