n8n-nodes-groq-speech-to-text

This is an n8n community node for transcribing audio using Groq's Whisper API. It's specifically designed to work seamlessly with Telnyx call recordings but supports any audio source.

Features

• Fast & Affordable: Uses Groq's ultra-fast Whisper models (whisper-large-v3-turbo and whisper-large-v3)
• Multiple Input Types:
  • Direct URL support (perfect for Telnyx recording_urls)
  • Binary data from previous nodes
• Two Operations:
  • Transcribe: Convert audio to text in original language
  • Translate: Convert audio to English text
• Flexible Options:
  • Multiple response formats (JSON, verbose JSON, text)
  • Language specification for improved accuracy
  • Custom prompts for context
  • Timestamp granularities (segment and word-level)
• Supported Audio Formats: MP3, MP4, WAV, M4A, FLAC, OGG, WebM, MPEG

Installation

Community Nodes (Recommended)

1. Go to Settings > Community Nodes in your n8n instance
2. Select Install
3. Enter n8n-nodes-groq-speech-to-text
4. Agree to the risks and click Install

Manual Installation

Navigate to your n8n installation and run:

npm install n8n-nodes-groq-speech-to-text

Restart n8n to load the node.

Prerequisites

• n8n instance (self-hosted or cloud)
• Groq API key from console.groq.com

Configuration

Credentials

1. In n8n, go to Credentials
2. Create new Groq API credentials
3. Enter your API key from Groq Console
4. Test and save

Node Parameters

Basic Settings

• Operation: Choose between Transcribe or Translate
• Input Type:
  • Audio URL: Provide a direct URL to the audio file
  • Binary Data: Use audio from a binary property
• Model:
  • whisper-large-v3-turbo: Faster, more affordable ($0.04/hour)
  • whisper-large-v3: Higher accuracy ($0.111/hour)

Optional Settings

• Language: Specify language code (e.g., 'en', 'es') for better accuracy
• Response Format: Choose output format (JSON, verbose JSON, or text)
• Additional Options:
  • Prompt: Guide transcription with context (max 224 tokens)
  • Temperature: Control randomness (0-1)
  • Timestamp Granularities: Get timestamps at segment/word level

Usage Examples

Example 1: Transcribe Telnyx Call Recordings

Perfect workflow for transcribing Telnyx call recordings:

Telnyx Webhook → Groq Speech-to-Text → Database/CRM

Telnyx Webhook Configuration:

• Listen for call.recording.saved events
• Extract recording_urls from webhook payload

Groq Speech-to-Text Node:

• Operation: Transcribe
• Input Type: Audio URL
• Audio URL: {{ $json.recording_urls.mp3 }} or {{ $json.recording_urls.wav }}
• Model: whisper-large-v3-turbo
• Response Format: verbose_json

The node will:

1. Download the audio from Telnyx
2. Send it to Groq for transcription
3. Return the full transcript with metadata

Example 2: Binary Data from File Upload

HTTP Request (get audio) → Groq Speech-to-Text → Process Text

Groq Speech-to-Text Node:

• Operation: Transcribe
• Input Type: Binary Data
• Binary Property: data
• Model: whisper-large-v3-turbo

Example 3: Translate Foreign Language Calls to English

Telnyx Webhook → Groq Speech-to-Text → Email/Slack

Groq Speech-to-Text Node:

• Operation: Translate
• Input Type: Audio URL
• Audio URL: {{ $json.recording_urls.mp3 }}
• Model: whisper-large-v3

Telnyx Integration Details

Webhook Event Structure

When Telnyx sends a call.recording.saved webhook, it includes:

{
  "recording_urls": {
    "mp3": "https://...",
    "wav": "https://..."
  },
  "recording_id": "...",
  "call_control_id": "...",
  "recording_started_at": "...",
  "recording_ended_at": "..."
}

You can directly use these URLs in the Groq Speech-to-Text node.

Recommended Workflow

1. Telnyx Webhook Trigger: Listen for call.recording.saved
2. Groq Speech-to-Text: Transcribe the recording
3. Post-Processing:
   • Store in database
   • Send to CRM
   • Analyze sentiment
   • Generate summary

Output Format

JSON Response Format

{
  "text": "The transcribed text..."
}

Verbose JSON Response Format

{
  "text": "The transcribed text...",
  "language": "en",
  "duration": 123.45,
  "segments": [
    {
      "id": 0,
      "start": 0.0,
      "end": 5.2,
      "text": "First segment..."
    }
  ],
  "metadata": {
    "model": "whisper-large-v3-turbo",
    "operation": "transcribe",
    "language": "en",
    "inputType": "url",
    "fileName": "recording.mp3"
  }
}

Pricing

Groq's Whisper API pricing (as of 2025):

• whisper-large-v3-turbo: $0.04 per audio hour
• whisper-large-v3: $0.111 per audio hour

Minimum billing is 10 seconds per request.

Rate Limits & File Size

• Free Tier: 25 MB max file size
• Dev Tier: 100 MB max file size
• Minimum audio length: 0.01 seconds

For large files, consider chunking audio before transcription.

Troubleshooting

Error: "Unable to locate package"

• Ensure you're using n8n version 0.200.0 or higher
• Try manual installation via npm

Error: "Invalid audio format"

• Supported formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, webm
• Check file is not corrupted
• Verify file size is within limits

Error: "Invalid API key"

• Verify your Groq API key is correct
• Check credentials are properly configured in n8n

Poor Transcription Quality

• Specify the language parameter
• Use whisper-large-v3 instead of turbo for better accuracy
• Add a context prompt to guide transcription
• Ensure audio quality is good (16KHz mono recommended)

Development

# Clone the repository
git clone <your-repo-url>
cd n8n-nodes-groq-speech-to-text

# Install dependencies
npm install

# Build the node
npm run build

# Development mode (watch for changes)
npm run dev

# Lint code
npm run lint

# Fix linting issues
npm run lintfix

# Format code
npm run format

Testing

To test the node locally:

1. Link the package to your n8n installation:

npm run build
npm link
cd ~/.n8n/nodes
npm link n8n-nodes-groq-speech-to-text

2. Restart n8n
3. The node should appear in the nodes panel

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Submit a pull request

License

MIT

Resources

• Groq Speech-to-Text Documentation
• Telnyx Voice API Documentation
• n8n Community Nodes Documentation

Support

For issues and questions:

• GitHub Issues: [Your repo URL]
• n8n Community Forum: n8n.community

Built with ❤️ for the n8n community