📚 Documentation
Last updated: 2026-02-08

Global Realtime Transcription (Beta)

TODO (New Screenshot): Global Realtime floating window + runtime state (App 2.0) Include: live transcript stream, model/language indicator, runtime state, and save controls. Suggested filename: global-realtime-window-v2-en.png

Scope

Global Realtime is a Beta workflow for system-wide audio capture and live transcription.

Capabilities include:

  • Cross-application audio capture
  • Dedicated realtime window output
  • Optional translation/save behavior

Use Cases

  • Continuous capture across app switching
  • Live monitoring and highlight extraction
  • System-level workflows where single-app capture is insufficient

Steps

  1. Launch Global Realtime Transcription (Beta) from Home.
  2. Select model, language, and save preferences.
  3. Start and confirm the realtime window is visible.
  4. Monitor transcript stream and runtime state.
  5. Stop and verify saved output/state.

Stability Tips

  • Start with a 5–10 minute smoke run before long sessions.
  • Save periodically for long-running captures.
  • If resource pressure is high, lower model size before increasing retries.

Real Scenario: Multi-app Meeting + Demo Capture

When browser meetings, local demos, and chat tools run together, single-app capture often misses critical audio. A safer operating pattern is:

  1. Run a short preflight capture before the session starts.
  2. Save checkpoint outputs at major meeting milestones.
  3. Move results into Note Workspace for final structuring and sharing.

Common Mistakes

  • Mistake 1: Starting long sessions without a smoke test.
    Fix: always validate permissions and window behavior first.
  • Mistake 2: Using Global Realtime by default for all tasks.
    Fix: if app-scoped capture is enough, use Realtime App Transcription.
  • Mistake 3: Ignoring runtime resource pressure.
    Fix: monitor CPU/GPU/memory and downgrade model tiers when needed.

FAQ

Q: Why does the realtime window not appear?
A: Check permissions, model readiness, and feature entitlement state.

Q: How is this different from Realtime App Transcription?
A: Global Realtime captures at system scope; Realtime App targets a selected app.

Q: Can it run for long sessions?
A: Yes, but you should monitor resource usage and periodically save results.

Limitations

  • Beta behavior/UI can change across versions.
  • Usually requires activated advanced entitlements.
  • Long-running sessions can consume significant CPU/GPU/memory.
  • Platform: Supported on Windows and macOS, with different system-audio permission flows and runtime stability patterns.
Whisper-Powered Live Transcription: Capture Speech from Mic, Apps & Media Files in Real Time

Contact us

Email
Copyright © 2026. Made by AudioNote, All rights reserved.