Prototype a local-first voice-note app
Start with a recording you can save, reopen, export, and delete without a network connection. Give every note a stable ID so a later transcript can point back to the original audio.
Start with one complete capture
Ask for microphone access when the user starts a recording. In a browser prototype, use a secure context and check the target device’s recording support. Show separate states for recording, saving, and saved. Display “saved” only after the local write completes.
Store the audio and its source record together in a local database such as IndexedDB. To support an offline reload, cache the app shell separately and test it after an online visit. Browser-managed data can be lost when the user clears site data, so give the user an export path.
Keep a source record
| Field | What it makes possible |
|---|---|
| Note ID | Update a local note and keep its identity through a retry. |
| Capture time and duration | Find the clip and tell whether it ended early. |
| Audio reference and MIME type | Open the original in a supported player. |
| Save result | Distinguish a completed local write from a failed capture. |
Run six release checks
- Deny microphone permission and confirm the app tells you how to retry.
- With a cached app shell, disconnect the network and save a short note.
- Reload offline and replay the note from storage.
- Simulate a storage write failure and check that the UI reports it.
- Export the audio and source record, then open both outside the app.
- Delete the note, reload, and verify that both local records are gone.
Add processing after capture is reliable
Keep transcription results linked to the note ID and store the processing state separately. If you later add an external service, let the user see which clip will leave the device and choose when to send it.
For a separate wearable-interface project exploring quick capture, follow MICRO’s development.