Audio Transcriber

A single-page HTML application that transcribes audio files using OpenAI’s Whisper API. Perfect for transcribing interviews, meetings, podcasts, and other audio content.

Features

• Drag & Drop Upload - Simply drag audio files into the browser
• File Picker - Traditional file selection interface
• Large File Support - Automatically chunks files over 25MB for processing
• Multiple Audio Formats - Supports MP3, WAV, M4A, FLAC, and more
• Visual Progress Tracking - Step-by-step progress indicator with time estimates
• Copy & Download - Copy transcription to clipboard or download as text file
• API Key Management - Securely stores OpenAI API key in localStorage
• Offline Ready - Self-contained HTML file, no external dependencies

Quick Start

1. Get OpenAI API Key
   • Visit OpenAI Platform
   • Create a new API key
   • Make sure you have credits available
2. Open the Tool
   • Open audio-transcriber.html in any modern web browser
   • Chrome, Firefox, Safari, and Edge are all supported
3. Add API Key
   • Enter your OpenAI API key when prompted
   • It’s stored securely in your browser’s localStorage
4. Upload & Transcribe
   • Drag an audio file onto the page, or use “Choose File”
   • Click “Transcribe” and wait for processing
   • Copy or download your results

Supported File Types

• Audio: MP3, WAV, FLAC, M4A, AAC, OGG
• Video: MP4, AVI, MOV, MKV (audio track will be extracted)
• Size Limit: 25MB per file (larger files are automatically chunked)

How It Works

Small Files (< 25MB)

1. Preparing - File validation and setup
2. Transcribing - Direct upload to OpenAI Whisper API
3. Complete - Results ready for copy/download

Large Files (> 25MB)

1. Preparing - File validation and setup
2. Analyzing Audio - Extract audio properties
3. Creating Chunks - Split into 5-minute segments
4. Transcribing - Process each chunk sequentially
5. Complete - Combine all chunks into final transcription

Technical Details

• Frontend: Vanilla JavaScript with React (CDN)
• Audio Processing: Web Audio API for chunking
• API: OpenAI Whisper-1 model
• Storage: localStorage for API key persistence
• Dependencies: None (fully self-contained)

Privacy & Security

• API Key: Stored locally in your browser only
• Files: Processed directly by OpenAI, not stored on any intermediate server
• No Tracking: No analytics or external requests (except to OpenAI)
• Local Processing: Audio chunking happens in your browser

Troubleshooting

“Invalid API Key”

• Verify your OpenAI API key is correct
• Check that you have available credits
• Try regenerating your API key

“File Too Large” (Old Error)

• This tool automatically handles large files
• Files are chunked into 5-minute segments
• Maximum practical limit is around 2GB

“Transcription Failed”

• Check your internet connection
• Verify the file is a valid audio/video format
• Try a shorter audio file to test your setup

Poor Quality Results

• Use higher quality audio files when possible
• Ensure clear speech with minimal background noise
• Consider preprocessing audio with noise reduction

Limitations

• Internet Required: Needs connection to OpenAI API
• OpenAI Credits: Costs apply based on audio duration
• Browser Memory: Very large files may use significant RAM during chunking
• URL Support: This version doesn’t support direct URL input (see MCP server for URL support)

API Costs

OpenAI Whisper API pricing (as of 2024):

• $0.006 per minute of audio transcribed
• Example: 1 hour of audio ≈ $0.36
• Chunked files are processed as separate requests but billed normally

Browser Requirements

• Modern Browser: Chrome 90+, Firefox 88+, Safari 14+, Edge 90+
• JavaScript: Must be enabled
• Features Used: File API, Web Audio API, Fetch API, localStorage

Files

• audio-transcriber.html - Complete application (single file)
• README.md - This documentation

---

Need URL Support? Check out the Audio Transcription MCP Server for YouTube and URL processing capabilities.