Overview
Groq provides ultra-fast inference for OpenAI’s Whisper models using their LPU (Language Processing Unit) architecture. This results in some of the fastest cloud transcription available. Key Benefits:- ⚡ Blazing fast - Sub-second transcription with Groq’s LPU technology
- 🎯 Whisper v3 accuracy - Latest OpenAI Whisper models
- 🌐 100+ languages - Full multilingual support
- 🎁 Free tier available - Generous free usage limits
Free Tier
What You Get
- Free tier with rate limits (exact limits vary, check Groq docs)
- All models included - Whisper v3, v3-turbo
- Same API for LLMs - Use for both transcription and AI enhancement
- No expiration - Free tier doesn’t expire
Getting Your Free API Key
1
Visit Groq Console
Go to console.groq.com in your browser.
2
Sign up for free
Click Sign Up and create an account with:
- Email and password
- Or sign in with Google/GitHub
3
Verify your email
Check your email for a verification link from Groq and click it to activate your account.
4
Create an API key
Once logged in:
- Go to API Keys section
- Click Create API Key
- Give it a name (e.g., “Stenox”)
- Copy the API key immediately (you can’t view it again!)
5
Add key to Stenox
- Open Stenox Settings
- Go to Transcription tab
- Select Groq as provider
- Paste your API key
- Choose your model (whisper-large-v3-turbo recommended)
- Click Save
One API key for everything: This same key works for Groq Whisper (transcription) and Groq LLMs (AI enhancement). Configure once, use for both!
Available Models
Groq offers multiple Whisper variants optimized for their LPU hardware:whisper-large-v3-turbo (Recommended)
Fastest and most efficient:- Latest Whisper v3 turbo variant
- Optimized for speed on Groq LPUs
- Sub-second processing for most recordings
- Best accuracy-to-speed ratio
whisper-large-v3
Maximum accuracy:- Full Whisper v3 model (largest size)
- Slightly slower than turbo
- Highest possible accuracy
- Best for: Critical transcription where accuracy is paramount
Language Support
Groq Whisper supports 100+ languages, same as OpenAI Whisper: Popular languages:- European: English, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Russian
- Asian: Chinese (Mandarin), Japanese, Korean, Hindi, Vietnamese, Thai, Indonesian
- Middle Eastern: Arabic, Hebrew, Turkish, Farsi
- Other: Ukrainian, Swedish, Norwegian, Finnish, and 70+ more
- Enable in Stenox profile settings
- Groq automatically identifies the language
- Works across all supported languages
Configuration in Stenox
1
Open Stenox Settings
Click Stenox menu bar icon → Settings
2
Go to Transcription tab
Select the Transcription tab
3
Select Groq
Choose Groq from the provider dropdown
4
Enter API key
Paste your API key from console.groq.com
5
Choose model
Select whisper-large-v3-turbo (recommended) or whisper-large-v3
6
Configure language
- English only: Select
English - Other languages: Select your preferred language
- Multiple languages: Select
Auto-detect
7
Save
Click Save to activate Groq Whisper
Features
LPU-Accelerated Processing
Groq’s Language Processing Units (LPUs) deliver exceptional speed:- Near-instant results - Sub-second for most recordings
- Consistent performance - No variability in processing time
- Scalable - Same speed regardless of usage volume
File-Based Transcription
Groq uses file-based transcription (not streaming):- Audio recorded completely, then sent
- Processed as a single file
- Results returned all at once
- Slightly different from streaming providers (DeepGram, AssemblyAI)
- You get the complete transcript at once (no partial results)
- Slightly higher latency than streaming (~1-2 seconds)
- Still very fast overall
Automatic Punctuation
Whisper models include automatic punctuation:- Periods, commas, question marks, exclamation points
- Capitalization of proper nouns
- Natural sentence formatting
Performance
Expected processing times with Groq Whisper:| Recording Length | Processing Time |
|---|---|
| 5 seconds | < 1 second |
| 10 seconds | 1-2 seconds |
| 30 seconds | 2-3 seconds |
| 60 seconds | 3-4 seconds |
Groq’s LPU architecture makes it one of the fastest cloud transcription options available.
Usage Tracking
Monitor your Groq usage in the console:- Visit console.groq.com
- Go to Usage or Billing section
- View:
- API calls made
- Rate limit status
- Current tier (free or paid)
Stenox does not track Groq usage. Check the Groq console for usage details and rate limits.
Privacy Considerations
What Groq sees:- Your audio is sent to Groq servers for processing
- Audio and transcripts may be logged for service improvement
- Data is encrypted in transit (HTTPS)
- Groq has data retention policies (review their docs)
- Use WhisperKit (local) for 100% offline transcription
- Or review Groq’s privacy policy for details
When to Use Groq Whisper
Speed is critical
Near-instant processing with Groq’s LPU technology.
Multilingual needs
Support for 100+ languages with auto-detection.
Free tier usage
Generous free tier perfect for regular use.
Whisper quality
Latest OpenAI Whisper v3 models with excellent accuracy.
Groq Ecosystem Integration
Use Groq for both transcription AND AI enhancement:1
Transcription with Groq Whisper
Convert speech to text using whisper-large-v3-turbo
2
Enhancement with Groq LLMs
Use Llama, Qwen, or DeepSeek R1 for grammar correction and formatting
3
Same API key
One API key, one provider, seamless integration
Troubleshooting
API key not working
API key not working
- Ensure you copied the complete API key (no spaces)
- Verify your Groq account is active (check email verification)
- Try creating a new API key in console.groq.com
- Check for typos or extra characters
Rate limit exceeded
Rate limit exceeded
- You’ve hit the free tier rate limit
- Wait a few minutes and try again
- Upgrade to paid tier for higher limits
- Use WhisperKit (local) as backup
Transcription fails
Transcription fails
- Check your internet connection
- Verify API key is valid in Groq console
- Ensure audio is being captured (check recording overlay)
- Try a shorter test recording
Wrong language transcribed
Wrong language transcribed
- Manually select language in Stenox profile (don’t use auto-detect)
- Ensure you’re speaking clearly in the selected language
- Try speaking longer phrases (short phrases confuse auto-detect)
Pricing After Free Tier
Groq offers competitive pay-as-you-go pricing when you exceed free tier limits:- Pricing varies by model - Check console.groq.com for current rates
- Generally very affordable - Competitive with other cloud providers
- No commitment - Pay only for what you use
Groq pricing is subject to change. Check their console for the most up-to-date pricing information.
Comparison: Groq vs Other Providers
| Feature | Groq Whisper | DeepGram | WhisperKit |
|---|---|---|---|
| Provider cost | Free (rate-limited) | $200 credits | No cost (local) |
| Speed | Very fast (1-2s) | Fastest (< 1s) | Medium (3-5s) |
| Accuracy | Excellent (Whisper v3) | Excellent (Nova-3) | Good (Whisper base/small) |
| Languages | 100+ | 30+ | 100+ |
| Privacy | Cloud | Cloud | 100% local |