How it works

From a press of a button to a searchable summary, entirely on infrastructure you control.

The pipeline

1. Record. A small ESP32-S3 device with a MEMS microphone captures audio to a microSD card. There is no length limit and no cloud round-trip to start recording.
2. Sync. Three independent paths. Bluetooth Low Energy for control and status. WiFi to the mobile app for fast transfer. And an independent WiFi upload straight to your storage whenever the device is on charge, with no phone involved.
3. Transcribe. Your server turns the audio into an accurate transcript using the transcription provider you configured, from a fully local model to a commercial API.
4. Summarise. An AI model writes a structured summary: an overview, the key points, and clear action items.
5. Export. Everything is available through a completely open REST API: the original audio, plain text, subtitles, Markdown, or JSON.

Three ways to sync

• Bluetooth for pairing, control and provisioning from the app.
• WiFi to the app for fast bulk transfer of recordings.
• Independent upload to your own storage when charging, so recordings arrive even if your phone is nowhere near.

Features