About Astrea

A desktop AI assistant that keeps you in control.

Astrea reads your project locally. When you make an AI request, only the context included in that request is sent through the gateway to the selected model service.

A desktop app that lives where your work already happens.

Astrea brings your files, editor, and AI conversation into one desktop workspace. Explore a project, work through a task, and inspect the result where the code lives.

Meet the team

Security as the baseline.

Good tools make the work easier to follow. Your project, the conversation, and the changes should be close enough to understand together.

Astrea does not upload or index your repository as a separate step. The context included in each model request is transmitted through the gateway; current model-service and deployment retention terms are described in the privacy policy.

Reads and edits your project on your machine
Your IDE token stays in your OS credential manager
Pick the model and cost profile for the task
Review proposed changes before applying them

Everything centers on your machine.

Astrea connects to the model you choose, reads your project locally, and applies your approved edits straight to your working tree.

01
Choose your approach
Start with the option that fits the task, then adjust as the work develops.
02
Work inside your project
Search your workspace and read the files you're actually touching, all on your disk.
03
Review it before applying
Review suggested edits in the git panel and apply them only when you say so.

What we guarantee

Local

Runs on your machine

Privacy questions answered with evidence

Private

Request context is sent only when an AI request is made

We do not publish a blanket no-retention promise

Honest

Usage and billing, side by side

Every prompt shows what it used and how it was billed

Yours

Astrea does not train its own models on requests

Pick the model for the task, from the ones we support