---
title: "Speech to Text: Convert a Recording to Text Free | ToolSugar"
description: "Convert a spoken recording into text in your browser with Whisper. Get TXT, SRT or VTT. Nothing is uploaded, and you get five free runs in any rolling 24 hours."
canonical: https://toolsugar.com/tools/speech-to-text
image: https://toolsugar.com/og.png
---

# Speech to text for recordings

Speech to Text turns a recorded voice, such as a memo, call or lecture, into written text. It works on an audio or video file you give it, not on live dictation. Whisper runs in your browser, so nothing is uploaded. Each run counts toward 5 free AI runs in any rolling 24 hours, with no sign up.

Runs on your device

## How it works

1. Drop a recording, or record one first with the Voice Recorder and drop the saved file here.
2. The recording stays in your browser. The first run downloads the Whisper model.
3. Set the Language to the spoken language, or leave it on automatic detection.
4. Press Transcribe, then copy the text or download it as TXT, SRT or VTT.

## Key facts

- **Input**: A recorded audio or video file, not live dictation
- **Engine**: Whisper runs in your browser after a one-time download of about 77 MB, and nothing is uploaded
- **Free runs**: 5 in any rolling 24 hours, no sign up
- **Outputs**: Plain text, SRT and VTT
- **Languages**: 22 to choose from, or automatic detection

## Common uses

- Turn a voice memo into notes you can search.
- Get a draft transcript of a phone interview before editing it.
- Make text from a recorded class so you can skim it.
- Check what was said in a short video without playing it through.

## Questions

### Can I speak into the microphone and get text live?

No. This tool transcribes a finished recording, it does not listen live. Record yourself first with the Voice Recorder, save the file, then drop it into Speech to Text.

### Is my voice recording uploaded?

No. The speech model runs in your browser, so the recording stays on your device. Only the model files download, once. Each run is counted with a hash of browser traits and a hash of your IP, not your audio.

### How accurate is the text?

Good on clear speech, weaker on Hindi and noisy audio, because the model is small. Speed depends on your device, and long files on weak devices are slow. Read the result before you rely on it.

### Does it add punctuation and speaker names?

It writes the text in paragraphs split at pauses. It does not label who is speaking, so for a conversation you add speaker names yourself.

[Read the Speech to Text guide](https://toolsugar.com/blog/how-to-transcribe-audio-to-text)

## Related tools

- [Transcribe to Text](https://toolsugar.com/tools/transcribe-audio-to-text)
- [Voice Recorder](https://toolsugar.com/tools/voice-recorder)
- [Auto Subtitles](https://toolsugar.com/tools/subtitle-generator)
- [Lecture to Notes](https://toolsugar.com/tools/lecture-to-notes)