---
title: "Audio to Text"
description: "Free audio to text transcription. Upload an MP3, WAV, or M4A and get an accurate transcript. Powered by a Pixeltable computed column calling a hosted Whisper model, so no GPU is required."
keywords:
  - speech to text free
  - transcribe audio
  - audio to text
  - audio transcription online
  - convert audio to text
  - mp3 to text
url: "https://www.pixeltable.com/tools/audio-transcriber"
---

# Audio to Text

Free audio to text transcription. Upload an MP3, WAV, or M4A and get an accurate transcript. Powered by a Pixeltable computed column calling a hosted Whisper model, so no GPU is required.

Free online tool by Pixeltable. Run it at https://www.pixeltable.com/tools/audio-transcriber (interactive, no sign-up).

## Features

- Whisper transcription
- Many languages
- Copy transcript
- Free

## FAQ

### What audio formats are supported?

MP3, WAV, M4A, and other common formats up to 50 MB.

### How accurate is the transcription?

It uses the Whisper model for high-quality transcription across many languages.

### Is it free?

Yes, completely free with no sign-up required.


## Fact sheet

Last updated 2026-09-23.

| Fact | Value |
| --- | --- |
| Modality | Audio |
| Inputs | Audio file |
| Outputs | Transcript text |
| Computed column | openai.transcriptions on the Audio column (model 'whisper-1') |
| What updates | A new or changed row recomputes only that row. Unchanged rows stay cached. |
| Canonical doc | [openai.transcriptions](https://docs.pixeltable.com/sdk/latest/openai#transcriptions) |

```python
import pixeltable as pxt
from pixeltable.functions.openai import transcriptions

TableModel = pxt.model_base()

class Clips(TableModel, name='clips'):
    audio: pxt.Audio

    transcript = transcriptions(audio, model='whisper-1').text
```