What audio formats work?

MP3, M4A (iPhone voice memos), WAV, OGG, OPUS (WhatsApp voice notes), AAC and FLAC — the built-in audio engine converts any of them before transcribing.

How long does a 1-hour recording take?

Roughly 10-30 minutes on a typical laptop with the Fast model, depending on your hardware. Machines with a modern GPU are several times faster. The page must stay open while it works.

Do you keep my recordings?

We never receive them. The transcription runs on your device — your recording and its transcript stay with you, full stop.

Does it tell me who is speaking?

No. The transcript is continuous text without speaker labels. For a two-person interview, the turn-taking is usually obvious from context; for a larger meeting, plan to mark speakers yourself while reviewing.

Should I convert the recording to MP3 before transcribing?

No. The built-in audio engine decodes M4A, OPUS, FLAC, WAV and the rest directly, so converting first only costs time and, for lossy formats, a little quality. Drop the original file in as it is.

Audio to Text — Free Private Transcription Online

Transcribe any audio recording to text in your browser. No upload, no signup, no per-minute limits.

Model

Spoken

Output

Timestamps [0:05]

🌍 The spoken language is detected automatically and the transcript comes out in that same language. Set it manually above if your audio is hard to hear, and use Output to translate into English.

⬇️ One-time download: 104 MB (Accurate model 73 MB + 31 MB audio engine, shared with the video tools) — only the first time: it stays saved in your browser, so next time it starts right away. (Translating or switching model needs its own file; a browser can clear this cache if storage runs very low.) Processing takes about 4-8 min per 10 minutes of audio on a typical laptop.

What exactly is downloaded — and is it safe?

• The speech model is OpenAI's Whisper, open source, converted to run in browsers. The audio engine is FFmpeg, the open-source tool behind most video software.
• They come from Hugging Face and unpkg — public repositories used by researchers and developers worldwide, not from us.
• They are not programs. Nothing is installed, no permission is asked, and your browser keeps them in its own cache exactly like it keeps a website's images.
• They can't read your files or touch anything outside this page. Remove them any time by clearing this site's data in your browser.
• Your audio never goes the other way: this site has no server receiving files.

Drop your audio here or click to choose

transcribed on your device by OpenAI's Whisper — your recording is not uploaded

ℹ️ Use a recording on your device that you own or have permission to process. This tool cannot fetch YouTube or social video links. Start with a short clip. Larger models can improve accuracy but require more memory and download data. Processing time varies by device, model and recording length. Always review the result.

Audio is the file type that most needs a transcript and is least suited to being uploaded: a meeting where colleagues were candid, an interview with a source, a lecture recorded without asking. This page runs OpenAI's Whisper speech model inside your browser, so the recording stays on your device while the text is produced, and there is no per-minute charge or cap. It takes the formats recorders actually produce, MP3, M4A from an iPhone voice memo, WAV, OGG, OPUS from WhatsApp, AAC and FLAC, and returns text to copy, save as TXT or export as subtitles.

Three models are offered. Fast is small and quick, fine for a clear voice in a quiet room. Accurate handles crosstalk, accents and background noise better. Best is the most demanding on memory and time and makes the fewest errors. On first use the audio engine and the chosen model are downloaded and cached when possible.

Treat the output as a draft. Whisper is good with ordinary speech and weaker with names, numbers, addresses and jargon, which are exactly what matters in a meeting, so read those back against the audio. It does not label who is speaking. Processing time depends entirely on your hardware: an hour-long recording on an older laptop is a job to leave running.

More sizes & options

Frequently asked questions

What audio formats work?: MP3, M4A (iPhone voice memos), WAV, OGG, OPUS (WhatsApp voice notes), AAC and FLAC — the built-in audio engine converts any of them before transcribing.
How long does a 1-hour recording take?: Roughly 10-30 minutes on a typical laptop with the Fast model, depending on your hardware. Machines with a modern GPU are several times faster. The page must stay open while it works.
Do you keep my recordings?: We never receive them. The transcription runs on your device — your recording and its transcript stay with you, full stop.
Does it tell me who is speaking?: No. The transcript is continuous text without speaker labels. For a two-person interview, the turn-taking is usually obvious from context; for a larger meeting, plan to mark speakers yourself while reviewing.
Should I convert the recording to MP3 before transcribing?: No. The built-in audio engine decodes M4A, OPUS, FLAC, WAV and the rest directly, so converting first only costs time and, for lossy formats, a little quality. Drop the original file in as it is.

Guides

Related tools

🔒 Private mode (the default): your recording never leaves your device. Instant mode is optional, you switch it on yourself, and it sends the audio to the browser's speech service.