Audio Note Quick Start

This guide targets desktop users on Windows and macOS.

home

TODO (Screenshot Replacement): Home overview (App 2.0 latest UI) Include: left navigation, four core entry cards (File/Realtime Mic/Realtime App/Link), top input area, and recent task list. Suggested filename: home-overview-v2-en.png

Quick Glossary

Whisper model: accuracy-first offline transcription models (usually heavier compute cost).
Realtime model (Sherpa): latency-first models for live feedback workflows.
Workspace: project-level data isolation to prevent cross-project mixing.
Watch Folder: automatic folder listener that queues new files for transcription.
Beta: features still under iteration; behavior and UI may change between versions.

Scope

Audio Note 2.0 focuses on desktop transcription and note workflows:

File transcription and batch transcription
Realtime transcription (microphone / app audio)
Recording and record-then-transcribe workflow
Link transcription (download + transcribe)
Watch folder automation
Workspace isolation, notes, and AI chat
Global realtime (Beta)
Dictation (coming soon, currently hidden by default)

Use Cases

Meeting capture and minutes
Learning notes from lectures and videos
Podcast/live replay processing
Team projects organized by workspace

Steps

Install from download page.
Open Settings and confirm model/download paths.
Download at least one model in Settings > Transcription.
Start one workflow from Home:
- Transcribe Files
- Realtime Transcription
- Realtime App Transcription
- Link Transcription
Review and edit results in the Note page.
Configure Watch if you need automatic folder processing.
Switch to the correct workspace before starting production tasks.

Real Scenario: Meeting Recap Onboarding

If you just finished a 60-minute meeting and need a shareable recap within 20 minutes, this path works well:

Drop the recording into File Transcription and run a first pass with Small/Medium.
In Note, fix names/numbers/terms first, then use AI Chat for action extraction.
Export Markdown for team sharing; add SRT if timeline review is needed.

Day-1 Checklist (Recommended)

Complete one file transcription and export TXT or SRT.
Complete one realtime microphone session to verify permissions/devices.
Use AI Chat once in Note to validate account/model readiness.
Review download concurrency in Settings to avoid unstable defaults.

Common Mistakes

Starting with the biggest model by default: often slower and less stable on modest hardware.
Running full batch without a smoke test: wrong parameters can create expensive rework.
Publishing AI output without review: manually verify names, numbers, and dates first.

FAQ

Q: Do I need to sign in first?
A: Not always. Basic local workflows are available without cloud features, while subscription features require sign-in.

Q: Why do I see fewer features than the docs?
A: Feature availability depends on account entitlements, OS permissions, and model installation state.

Q: How do I enable Beta features?
A: Beta features are controlled by app version/channel and feature flags; not-yet-released features remain hidden.

Limitations

Desktop only (Windows/macOS).
Some advanced workflows require activated subscription features.
Beta features can change quickly across versions.
Dictation is not generally available yet in current release channels.
Large models and high concurrency require stronger hardware resources.