How it works

Wassender includes an open speech-recognition model (Whisper) that runs inside Chrome. When you click Transcribe on a voice note, the audio is decoded in your WhatsApp tab and turned into text on your computer.

The first time, Chrome downloads the model once (about 40 MB for Fast). After that it works from the browser cache. Choose Better or Accurate for harder audio.

Tell it the language

Short clips are easier to transcribe correctly when the language is set. Pick Arabic, English or Portuguese, or leave it on automatic.

Step by step

  1. Open a chat with a voice note on WhatsApp Web.
  2. Click the small Transcribe chip next to the voice note.
  3. The first time, confirm the one-time model download.
  4. Read the text under the voice note. Transcripts are remembered in this browser.

Plan: Free.

Frequently asked questions

Is my audio sent to a server or an AI service?

No. The model runs locally in Chrome. Wassender never receives your audio or the transcript.

How accurate is it?

Clear speech in a quiet place transcribes well. Noisy clips and heavy dialects are harder; the Accurate model helps, but read important details twice.

Does it work offline?

After the model has been downloaded once, transcription itself runs on your computer.

Keep reading