# Speech to Text

> Transcribe audio to text in your browser. Upload a file or record from your microphone — a small local model (~40 MB) does the work. Nothing is uploaded.

URL: https://uttir.com/speech-to-text
Categories: media-tools, ai-tools
Privacy: Runs entirely on your device. The audio and the model stay on your device — nothing is uploaded. The first run downloads a small local model (~40 MB); after that, transcription works offline.

## About

This tool transcribes audio to text using a small AI model that runs in your browser. The audio and the model never leave your device.

The first run downloads the model (~40 MB). After that, transcription works offline. You can upload an audio file or record from your microphone.

Output is shown as text with timestamps. You can copy the text, download as .txt, or download as .srt subtitles (compatible with YouTube, VLC, and most video editors).

Everything — the audio, the model, and the result — stays on your device. There is no upload, no account, and no analytics on the audio itself.

## How to use

1. **Pick an audio source** — Choose an audio file (mp3, m4a, wav, webm, ogg) or click Record to capture from your microphone.
2. **Download the model on first use** — The first run downloads a small local model (~40 MB). After that, transcription works offline.
3. **Click Transcribe** — Inference is local. A one-minute clip typically finishes in a few seconds on a modern laptop. The result appears below with timestamps.
4. **Copy or download** — Use Copy for plain text, or download as .txt or .srt. The .srt format works with most video editors and YouTube for subtitles.

## FAQ

### How accurate is the transcription?

For clear English speech with a decent microphone, accuracy is high. For background noise, strong accents, or low-quality audio, expect more errors. The model is the smallest Whisper variant (tiny) for fast download — a larger model would be more accurate but much bigger.

### What languages are supported?

The model auto-detects the language and supports dozens of languages, with English being the strongest. The first few seconds of audio are used to detect the language.

### Does my audio leave my device?

No. The audio and the model both run in your browser. There is no server step. The microphone stream is local to your tab.

### Is the model downloaded every time?

No. The first run downloads it to your browser cache; subsequent runs use the cached copy and work offline. You can remove the model from your device at any time with the small button on the page.

### What is the .srt format?

SRT (SubRip) is the most widely supported subtitle format. Each entry has a sequence number, a timecode range, and the spoken text. YouTube, VLC, Premiere Pro, and Final Cut Pro all accept it.

## Related tools

- [Text Summarizer](https://uttir.com/text-summarizer) — Summarize long articles, blog posts, reports, or emails — entirely in your browser. Nothing is uploaded. The model runs locally on your device.
- [Sentiment Analyzer](https://uttir.com/sentiment-analyzer) — Detect the sentiment of any text — positive, negative, or neutral — using a local model. Perfect for content review, social posts, or feedback triage. Nothing is uploaded.
- [Text Similarity](https://uttir.com/text-similarity) — Compare two pieces of text and get a 0–1 similarity score plus a verdict (near-duplicate, similar, related, unrelated). Runs in your browser, nothing is uploaded.
- [Background Remover](https://uttir.com/background-remover) — Remove the background from any image — entirely in your browser. The first run downloads a small local model (~25 MB); after that it runs offline and nothing is uploaded.

---

For the full HTML page with the live tool, visit https://uttir.com/speech-to-text.
This file is the markdown rendering at https://uttir.com/speech-to-text.md. See https://uttir.com/llms.txt for a site-wide summary.