Voice Input

Voice input lets you speak your thoughts naturally. ThoughtFree transcribes your speech in real time using Deepgram’s Nova-3 model.

How to Record

1. Tap the FAB (+ button) to open the capture sheet
2. Tap the microphone icon
3. Start speaking — you’ll see your words appear as you talk
4. Tap Stop when finished
5. Review the transcription, edit if needed, and save

Tip

You can pause and resume recording. The transcription continues seamlessly from where you left off.

Live vs Batch Transcription

ThoughtFree uses live streaming transcription by default:

Mode	How it works	When it’s used
Live	Words appear in real time via WebSocket	Default — best experience
Batch	Audio is sent after recording completes	Fallback when WebSocket unavailable

The app automatically falls back to batch mode if the live connection drops. You don’t need to configure this.

Voice Settings

Customize voice capture in Settings > Voice:

Speed

Controls how quickly the transcriber finalizes words (utterance end detection):

Setting	Delay	Best for
Fast	1000ms	Quick notes, short phrases
Balanced	1200ms	General use (default)
Patient	1500ms	Thoughtful speech with longer pauses

Sensitivity

Adjusts how aggressively the transcriber detects speech phrases. Higher sensitivity picks up quieter speech but may also capture background noise.

Language

Select your primary language for recognition. This helps the model optimize accuracy for your language.

Tips for Better Transcription

1. Speak clearly at a natural pace — no need to slow down
2. Minimize background noise — the model handles some noise but quiet environments work best
3. Use the Patient speed if you pause often to think between sentences
4. Review before saving — transcription is very accurate but not perfect

Voice Minutes

Voice capture usage varies by plan:

Plan	Voice Minutes
Free	5 min/month
Pro	1,000 min/month

Minutes reset monthly. You can check your usage in Settings > Plan & Billing.