Transcribe Audio to Text Online

Transcribe spoken audio locally into TXT and SRT subtitles with Whisper Tiny.

Import your audio files

0 file
Speech to Text — Settings apply to every loaded file
Settings apply to every loaded file. Processing stays on your device.?

The higher-quality model requires more memory and a larger first download.

Whisper Tiny runs locally after a one-time model download (may be large). Short recordings are recommended. Outputs: TXT + SRT. First transcription may be slow.

Browser must support local audio decoding (typically MP3, WAV, M4A or OGG).

Transcription currently supports recordings up to 10 minutes per file. This is a processing limit, not an upload-list limit.

Add audio files to begin.
Processing progress 0%

Common uses and practical advice

Transcribe spoken audio locally with a downloadable speech recognition model. Review the text and export a transcript or SRT subtitles.

Transcribe interviews, voice memos or short lectures, then check names and punctuation manually. Accuracy depends on noise, accent and device performance.

Help & tips

How to transcribe audio to text

  1. Upload a short recording of spoken language.
  2. Choose the language and Whisper model quality.
  3. Transcribe and download TXT or timestamped SRT subtitles.

Example use cases

  • Turn interview speech into editable notes.

Limitations to consider

  • A first-use model download is required; processing time depends on device.
  • A recording is limited to 10 minutes for this browser version.
  • Speech recognition may omit words or invent text; proofread before publishing.

Questions and practical tips

Is the audio uploaded to a server?

No. Decoding and inference happen locally in a browser worker; model and JavaScript resources are fetched from third-party hosts.

Why are subtitles approximate?

Speech recognition timestamps are estimates, especially near chunk boundaries. Review the SRT before use.

Processing and file safety

This tool processes files in your browser. Loading the page, libraries or AI models can still require network connections. Keep a copy of your original files and check the downloaded result before deleting them. Large files may require more memory and processing time than your device can provide.

Explore tools and the formats guide

Go to homepage