n8n-nodes-audio-split-v2

Professional audio splitting node for n8n - Perfect for processing large audio files with OpenAI Whisper and other transcription services

This is an n8n community node that splits audio files into configurable chunks using ffmpeg. Designed specifically for n8n 2.0+ compatibility.

🎯 Why This Node?

• OpenAI Whisper Integration: Whisper API has a 25 MB file size limit - this node splits larger files automatically
• Configurable Chunks: Set custom chunk duration (default: 10 minutes)
• Smart Processing: Each chunk is returned as a separate n8n item for parallel processing
• Production Ready: Handles large files efficiently with automatic cleanup

✨ Features

• ✅ n8n 2.0+ Compatible - Works with the latest n8n binary data system
• ✅ Configurable Duration - Split by time (seconds)
• ✅ Multiple Outputs - Each chunk as separate workflow item
• ✅ Automatic Cleanup - No temporary file buildup
• ✅ Error Handling - Robust error reporting
• ✅ ffmpeg Powered - Fast, reliable audio processing

📋 Prerequisites

• n8n version 1.0.0 or higher (optimized for n8n 2.0+)
• ffmpeg installed in your n8n environment

🚀 Installation

In n8n (Recommended)

1. Go to Settings → Community Nodes
2. Click Install
3. Enter: n8n-nodes-audio-split-v2
4. Click Install

Manual Installation

# In your n8n installation directory
npm install n8n-nodes-audio-split-v2

# Restart n8n

Docker Installation

# If ffmpeg is not installed in your n8n container:
docker exec -it n8n apk add ffmpeg

# Install the node
docker exec -it n8n npm install -g n8n-nodes-audio-split-v2

# Restart container
docker restart n8n

🎬 Usage

Basic Workflow Example

Form Trigger (File Upload)
    ↓
Audio Split
    ↓
OpenAI (Whisper Transcription)
    ↓
Aggregate Results

Node Configuration

Parameter	Type	Default	Description
Input Binary Field	string	data	Name of the binary property containing the audio file
Chunk Duration	number	600	Duration of each chunk in seconds (600 = 10 minutes)
Output Prefix	string	chunk	Prefix for output chunk filenames

Example: Transcribe Large Audio File

// Workflow Setup:
// 1. Webhook/Form receives audio file
// 2. Audio Split (600 second chunks)
// 3. OpenAI Whisper processes each chunk
// 4. Aggregate node combines transcriptions

// Audio Split Configuration:
{
  "Input Binary Field": "data",
  "Chunk Duration": 600,     // 10 minutes
  "Output Prefix": "chunk"
}

// Output (per chunk):
{
  "json": {
    "fileName": "chunk_001.m4a",
    "originalFile": "interview.m4a",
    "chunkIndex": 1
  },
  "binary": {
    "data": {
      "fileName": "chunk_001.m4a",
      "mimeType": "audio/x-m4a",
      "data": "..."
    }
  }
}

🔧 Configuration Examples

For Whisper API (25 MB limit)

// Recommended: 10-minute chunks for most audio
{
  "Chunk Duration": 600
}

For Large Files (>1 hour)

// Smaller chunks for parallel processing
{
  "Chunk Duration": 300  // 5 minutes
}

For High-Quality Audio

// Longer chunks if file size allows
{
  "Chunk Duration": 900  // 15 minutes
}

📊 Real-World Use Cases

1. Podcast Transcription

File Upload → Audio Split (10 min) → Whisper API → Combine Transcripts → Save to Database

2. Meeting Recording Processing

Google Drive Trigger → Audio Split → Multiple AI Services (Whisper, Sentiment Analysis) → Report

3. Interview Analysis

Form Upload → Audio Split → Whisper Transcription → GPT-4 Summary → Email Results

🐛 Troubleshooting

Node doesn't appear in n8n

Solution:

# Check if installed
npm list n8n-nodes-audio-split-v2

# Restart n8n
docker restart n8n  # or your restart method

# Check n8n logs
docker logs n8n | grep audio-split

"ffmpeg not found" Error

Solution:

# For Docker/Alpine Linux
docker exec -it n8n apk add ffmpeg

# For Ubuntu/Debian
apt-get install ffmpeg

# Restart n8n
docker restart n8n

Chunks are empty or corrupted

Possible causes:

• Input file is corrupted
• Binary property name is incorrect
• Insufficient disk space

Solution:

1. Verify the input file plays correctly
2. Check the binary property name matches your input
3. Verify disk space: df -h

"The item has no binary field 'data'"

Solution:
The binary property name doesn't match. Check what your previous node outputs:

1. Click on the previous node
2. Look at the Binary tab
3. Use that property name in "Input Binary Field"

🔍 How It Works

1. Receives Binary Data: Gets audio file from previous node
2. Creates Temp Directory: Stores file temporarily
3. Runs ffmpeg: Splits audio using segment filter
4. Returns Chunks: Each chunk as separate workflow item
5. Cleanup: Automatically removes temp files

🤝 Contributing

Contributions are welcome! This is an open-source project.

Development Setup

# Clone the repository
git clone https://github.com/YOUR_USERNAME/n8n-nodes-audio-split-v2

# Install dependencies
npm install

# Build
npm run build

# Test locally
cp -r dist ~/.n8n/custom/n8n-nodes-audio-split-v2

Reporting Issues

Found a bug? Open an issue

📝 Version History

1.0.0 (2026-01-19)

• ✨ Initial release
• ✅ n8n 2.0+ compatibility
• ✅ Configurable chunk duration
• ✅ Multiple output support
• ✅ Automatic cleanup

📄 License

MIT License - feel free to use in your projects!

👨‍💻 Author

Leon @ ADYZEN

• Company: ADYZEN - AI Consulting & Automation
• Specialization: n8n workflows, RAG systems, process automation

🙏 Acknowledgments

• Built for the n8n community
• Powered by ffmpeg
• Designed for OpenAI Whisper integration

📚 Related Resources

• n8n Documentation
• n8n Community Nodes Guide
• OpenAI Whisper API Docs

---

⭐ If this node helped you, please star the repository!

Made with ❤️ for the n8n community