Raw data and computation
Raw matrices are handled locally by R/Python scientific environments. The protected Core does not need to receive the raw dataset.
Scientific-data privacy model
Your scientific data stay on your computer.
Local-first is an architectural choice: raw matrices and large scientific datasets stay in your local workspace, where scientific computation takes place.
On your computerLOCAL WORKSPACE
01
Your scientific data
Raw matrices & processed inputs
02
scRNA-agent workspace
Your question, context & decisions
03
R / Python analysis
Local scientific computation
04
Project outputs
Figures, tables, reports & provenance
Raw matrices stay on your computer. Scientific computation runs locally.
YOUR MODEL ACCESS
Your own AI provider supports reasoning and execution. Prompts and shared project context may leave your computer and are subject to that provider’s policies.
Understand each part of the boundary.
Project files and outputs
Local project files, figures, tables, and reports remain under your control. You choose how to store, back up, and share them.
Protected Core services
The Core provides workflow guidance, scientific skills, account and entitlement services, and updates. These connected services are distinct from raw-data computation.
Your AI/model provider
Your chosen provider supports reasoning and execution. Prompts and shared project context may leave your computer; local computation does not mean all AI interaction is offline.
Make informed choices about shared context.
Review your model provider’s policies and your institution’s requirements before sharing sensitive information. Do not include identifiable or restricted information in prompts unless you have an appropriate basis to do so.
This page explains the product’s data-handling architecture. It is not a blanket guarantee about every external provider or a complete legal privacy policy. Detailed service data disclosures will be published with the application release.
Ask about data handling