Supported Platforms
ElevenLabs
Capturing AI-generated voice and audio from ElevenLabs
ElevenLabs
ElevenLabs is a leading AI voice platform offering text-to-speech, voice cloning, and AI audio generation. Known for highly realistic voices and emotional range, it's widely used for voiceovers, narration, and character voices.
Platform Overview
- Provider: ElevenLabs
- Specialty: Text-to-speech, voice cloning, audio generation
- Access: Free tier available, paid plans for commercial use
- Supported Output Types: Audio, Voice, Narration
Lazer Detection
URL Patterns
The extension detects ElevenLabs on:
https://elevenlabs.iohttps://elevenlabs.io/speech-synthesishttps://elevenlabs.io/voice-labhttps://elevenlabs.io/audio
What Is Auto-Detected
When you generate with ElevenLabs, Lazer extracts:
- Prompt - The text to be spoken (script input)
- Voice - Selected voice name (e.g., "Rachel", "Adam")
- Model - TTS model version (e.g., "Eleven Multilingual v2")
- Settings - Stability, similarity, style sliders
- Output URL - Audio file URL
- Thumbnail - Waveform visualization (if available)
Detection Quality
- Prompt extraction: High - Dedicated text input area
- Settings extraction: Medium - Some settings in dropdowns
- Output extraction: High - Reliable audio URL detection
Capture Workflow
Text-to-Speech Workflow
- Open ElevenLabs Speech Synthesis page
- Select a voice from the voice library
- Enter your script in the text area
- Adjust settings (stability, similarity, style)
- Click "Generate"
- Wait for audio generation (usually 2-5 seconds)
- Audio player appears with waveform
- Open Lazer extension
- Detection banner shows "ElevenLabs"
- Click "Auto Fill"
- Review fields (especially voice name)
- Click "Save Capture"
Voice Cloning Workflow
- Upload voice samples to Voice Lab
- Train custom voice
- Select cloned voice
- Generate speech as above
- Capture with Lazer
Long-Form Narration
For audiobook or long-form content:
- Use Projects feature
- Upload script
- Generate chapter by chapter
- Capture each chapter separately
- Tag with chapter numbers for organization
Capture Tips
Best Practices
- Tag by voice - Use voice name as tag (e.g., "rachel", "adam")
- Note emotion - Include emotional direction in prompt field
- Check licensing - Free tier is non-commercial only
- Include script - Full script text is the "prompt"
Waveform Thumbnails
ElevenLabs shows waveform visualizations:
- Detector attempts to capture waveform image
- Useful for visual reference in web app
- Falls back to null if not available
Common Issues
Voice name not detected:
- Voice selector may be a dropdown
- Manually note voice name in title or tags
- Critical for tracking which voice was used
Settings not captured:
- Stability, similarity, style are advanced settings
- Not always visible in UI
- Manually add to metadata JSON if needed
Long text truncated:
- ElevenLabs has character limits per generation
- For long form, capture in chunks
- Include sequence info in title (e.g., "Chapter 1")
Rights and Licensing
Free vs Paid Plans
ElevenLabs has strict licensing by plan:
Free Plan
- Non-commercial use only
- 10,000 characters per month
- Attribution required
- Cannot use in commercial projects
Creator/Pro Plans
- Full commercial license
- Higher character limits
- No attribution required
- Use in client projects, YouTube, podcasts, etc.
Attribution Requirements
Free plan requires:
- Credit ElevenLabs in your project
- Format: "Voice generated by ElevenLabs"
- See ElevenLabs attribution guidelines
Voice Library Licenses
- Pre-made voices: Licensed for use per your plan
- Professional Voice Library: Additional licensing terms
- Cloned voices: You retain rights to your voice
Lazer Rights State
The extension sets rightsState to:
- RESTRICTED - If free plan detected (requires manual confirmation)
- LICENSED - If paid plan (inferred from API)
- UNKNOWN - Default
Always verify your ElevenLabs plan before commercial use. Set rights state manually in web app.
Known Limitations
Voice Selector UI
Voice selection UI varies:
- Dropdown, modal, or sidebar
- Detector uses text content from active voice
- May miss voice name if UI changes
Settings Hidden
Advanced settings (stability, similarity) may be:
- In expandable panel
- Not visible until clicked
- Manually capture if needed for reproducibility
Audio Format
ElevenLabs outputs MP3 by default:
- Detector captures audio URL as-is
- Format is in metadata if detectable
- Check audio player
srcattribute
Metadata Fields
ElevenLabs-specific metadata:
{
"voiceName": "Rachel",
"model": "Eleven Multilingual v2",
"stability": 0.5,
"similarity": 0.75,
"style": 0.0,
"format": "mp3",
"duration": 45.2,
"platform": "elevenlabs"
}Example Capture
A typical ElevenLabs capture in Lazer:
- Platform: elevenlabs
- Asset Type: VOICE
- Prompt: "Welcome to the world of tomorrow, where anything is possible."
- Voice: Rachel
- Output URL:
https://storage.elevenlabs.io/audio_abc123.mp3 - Thumbnail:
https://storage.elevenlabs.io/waveform_abc123.png - Metadata:
{"voiceName": "Rachel", "model": "Eleven Multilingual v2", "duration": 4.5} - Status: GENERATED
- Rights State: RESTRICTED (free) or LICENSED (paid)
Troubleshooting
Detection Not Working
- Verify you're on the Speech Synthesis page
- Wait for audio generation to complete
- Click "Refresh Detection"
- Grant host permissions for
elevenlabs.io
Voice Name Missing
- Detector looks for voice badge or selector text
- If missing, manually add voice name to title field
- Critical for tracking which voice was used
Audio URL Expired
- ElevenLabs may use temporary URLs
- Extension syncs immediately
- Verify successful sync in Queue mode
Settings Not Captured
- Stability, similarity, style are optional
- Manually document in metadata if needed for reproduction
- Most important for voice matching across generations
Commercial License Unclear
- Free plan: Non-commercial only, attribution required
- Creator+: Commercial use allowed
- Review ElevenLabs pricing page for current terms
- Set rights state to RESTRICTED if on free plan