whisper.cpp : WASM example

Free Online Whisper.cpp - AI Speech-to-Text Transcription whisper.cpp running fully in your browser

Usage Instructions:

Load a Model: Choose from the models below or upload your own ggml file
Upload Audio: Select any supported audio file (see formats below)
Transcribe: Click "Transcribe" to convert speech to text
Download: Save your transcription as a text file with timestamps

Supported Audio Formats:

.wav .mp3 .mp4 .m4a .ogg .webm .flac .aac

Note: All audio formats are automatically converted to the required format for processing.

Available AI Models & Their Performances:

Model	Size	Speed	Quality	Best For
tiny.en	75 MB	Fastest	Basic	Quick testing, clear speech
base.en	142 MB	Fast	Good	General use, good balance
small.en	466 MB	Moderate	High	Professional transcription
medium/large	515MB+	Slow	Best	Highest accuracy needed
base (multilingual)	142 MB	Good	Good	99+ languages supported
large (multilingual)	1030 MB	Slowest	Best	Best quality, all languages

Recommendation: Start with base.en for best speed/quality balance

Status: Whisper not loaded

Whisper models:

Quantized models:

Input: File Microphone

Upload Audio File:
Supports: WAV, MP3, FLAC, M4A, AAC, OGG, WMA, WEBM | Max recommended: 100MB | Max length: ~30 minutes

Click on the "Transcribe" button to start the transcription

Note that the computation is quite heavy and may take a few seconds to complete. The transcription results will be displayed in the text area below.

Language:

Threads: 8

About This Service

Privacy & Security

All processing happens in your browser
No files uploaded to servers
Your audio never leaves your device
No registration or login required

Performance Tips

Use base.en for best balance
Works best on desktop browsers
Requires WASM SIMD support
Longer files = more processing time

Free & Open Source

Completely free to use
Based on OpenAI Whisper
Powered by whisper.cpp
Open source on GitHub