Getting started
New here? This page orients you: the one idea that shapes the whole system, and the fastest path through the docs for whatever you came to do. For what each piece is, see the Overview; for how it all works, see Architecture.
The one idea
Audio flows from a device in someone's hand to a finished report on the web: captured on hardware, synced to the cloud, transcribed and analysed by an AI model, then presented to the clinician. Almost every design decision follows from a single guarantee.
A recording is never treated as safely delivered until the cloud has verifiably stored it. A device keeps its own copy of audio until storage is confirmed byte-for-byte, and the long-running AI step lives in a process that can take as long as it needs — never on a request that might time out mid-transcription.
Find your path
Reading paths by task
| If you're… | Start with | Then read |
|---|---|---|
| New to the project | Overview → Architecture | The component guide for your area |
| Building recorder / pendant firmware | Recorder · Pendant | Hardware testing · Firmware release |
| Working on the backend or pipeline | Backend pipeline | Device API · Data model |
| Building the mobile or web app | Mobile app · Web app | BLE protocol |
| Operating or testing devices | Hardware testing | The sate CLI · Troubleshooting |
The parts at a glance
Each of these is versioned and released independently. The component guides cover them one by one; the Version log tracks what changed in each release.