Browser tool · Local processing

Audio to text

Transcribe an audio file locally with a downloadable Whisper model.

Transcribe audio to text in your browser

Run Whisper Tiny on this device

Transcribe an audio file locally with a downloadable Whisper model.

Copy or download the transcript as TXT

Supported formats and codecs depend on your browser and source file.

Your audio is processed on this device. The Whisper model downloads in your browser.

Transcript

Speech-to-text example

Try a short speech sample

This 11-second English speech clip is from the public OpenAI Whisper test assets. Play or download it, then choose the file above to try this transcription tool.

Download WAV sample

Sample source: OpenAI Whisper · MIT license

Reference page sample output

And so my fellow Americans
Ask not
What your country can do for you as what you can do for your country

This is the reference page’s recorded Tiny + WebGPU output and is uncorrected. This tool uses Tiny + WASM, so results may differ. Review names, numbers, and punctuation.

What this tool supports

The first run downloads the Whisper Tiny model. Audio decoding and recognition run in this browser, while the model and code still require network access. This tool currently exports plain TXT only; it does not provide speaker labels or transcript proofreading.

If transcription fails, try a shorter MP3, WAV, or M4A file, confirm its audio track plays, and check that the device has enough memory. Results vary with language, noise, accent, and device performance.

How to use this tool

  1. 1

    Choose an audio file from your device. Try the linked WAV speech sample if you want to check the workflow first.

  2. 2

    On first use, wait for the Whisper Tiny model and browser audio decoder to load.

  3. 3

    Select Transcribe on this device, review or edit the transcript, then copy it or download the TXT file.

Why use this tool

Transcribe an audio file to text in the browser without sending the media to a server. The model is downloaded to your browser on first use and recognition then runs with WebAssembly on your device.

Frequently asked questions

How do I transcribe audio to text?

Choose a local audio file, wait for the model to load if this is the first run, then select Transcribe on this device. Review the generated text before copying or downloading the TXT file.

Can I transcribe an audio file for free online?

This browser tool does not charge for transcription. Your device must download the Whisper Tiny model, and processing speed depends on its CPU, memory, and the length of the audio.

Why is the first transcription slower?

The browser needs to download and initialize the Whisper Tiny model before it can process speech. Later visits may use the browser cache, but the cache can be cleared.

Which file types and languages work?

The file picker accepts audio files, but successful decoding depends on the audio codec supported by the browser decoder. Whisper Tiny is multilingual; recognition quality depends on the spoken language and recording quality.

Why is my audio transcript inaccurate or empty?

Noise, music, silence, overlapping voices, uncommon names, and unsupported codecs can affect the result. Try a short, clear audio file and review the transcript manually.

Is speech recognition completely offline?

The selected audio is processed on this device, but the web app and Whisper model need network access when they are first loaded.