Cloud Speech-to-Text
GetThatQuick supports two STT modes:
| Mode | Requires | Works Offline | Best For |
|---|---|---|---|
| Local (Vosk) | Downloaded model (40 MB+) | ✅ Yes | Privacy, no internet |
| Cloud (Groq/OpenAI) | API key | ❌ No | Low-RAM devices, fast setup |
Groq Whisper (Recommended — Free)
Groq provides free access to whisper-large-v3-turbo — extremely fast, accurate, and OpenAI-compatible.
Free tier limits:
- 8 hours of audio per day
- 20 requests per minute
- No credit card required
Setup
- Get a free API key at console.groq.com → API Keys
- In GetThatQuick: Settings → Voice / STT
- Select Groq Whisper as provider
- Paste your API key (
gsk_...) - Pick a model (
whisper-large-v3-turbois recommended)
Usage
- Click the mic button in the chat input (shows a small blue dot when cloud STT is active)
- Speak — a waveform animation appears showing your audio level
- Click the mic again (or the MicOff button) to stop recording
- A brief "Transcribing…" appears, then your words fill the input
- Edit if needed, then send
OpenAI Whisper
If you already have an OpenAI API key:
- Settings → Voice / STT
- Select OpenAI Whisper
- Paste your OpenAI API key (
sk-...) - Select a model:
whisper-1— classic, cheapestgpt-4o-mini-transcribe— faster, smartergpt-4o-transcribe— best accuracy
Pricing: $0.006/minute ($0.36/hour)
Switching Between Local and Cloud
You can switch at any time without losing data:
- Settings → Voice / STT → STT Provider
- Click Local (Vosk), Groq Whisper, or OpenAI Whisper
- The preference is saved immediately
The mic button adapts its tooltip to show the active mode.
Troubleshooting
| Problem | Fix |
|---|---|
| "No API key configured" | Add your key in Settings → Voice |
| "Audio recording too short" | Hold the mic button at least 1–2 seconds |
| "STT provider error (401)" | API key is invalid or expired |
| "STT provider error (429)" | Rate limit hit — wait a moment or upgrade plan |
| Transcription is empty | Try speaking more clearly or closer to the mic |
| No waveform visible | Browser may not support Web Audio API (use Chrome/Firefox) |