Turn speech into text with AI
This tool writes down what is said in an audio or video file: a meeting, an interview, a podcast, a lecture, a voice memo or a video for social media. It uses Whisper, an AI model that recognizes 99 languages and adds the punctuation by itself. Every sentence keeps its time in the recording: click one to hear it, then download the transcript as plain text, or as SRT and VTT subtitles for your videos. You see the first lines for free before paying anything.
How to transcribe a recording
- Drop your audio or video file in the box. The first lines are ready in a few seconds.
- Check the language and the first lines. Change the language if it was not detected right.
- Click Get the full transcript, pay €1.99 once, then copy the text or download TXT, SRT or VTT.
What to use it for
- Meetings and interviews: the written record of a call, to quote, search or summarize.
- Subtitles and captions: an SRT or VTT file for YouTube, TikTok, Instagram or a course video.
- Podcasts: show notes, a blog article or a searchable page from an episode.
- Lectures and voice memos: notes you can read instead of listening again.
- Audio in another language: the written text, easier to translate than the sound.
Good to know
- Files up to one hour (and 500 MB). For a longer recording, cut it with the audio cutter first.
- The whole recording is transcribed after the payment, in parts of two minutes: count about one minute for 10 minutes of audio. Keep the tab open.
- One language per recording: the language found at the start (or the one you choose) is used for the whole file. A recording that switches language often is transcribed less well.
- Speakers are not named: the transcript does not say who is talking. The text keeps a new paragraph after each pause.
- If something fails after the payment, "Try again" goes on with the same payment. If it keeps failing, contact us with your receipt and we refund you.
Frequently asked questions
How do I convert an audio file to text?
Drop the file (MP3, WAV, M4A, OGG, FLAC, or a video: MP4, MOV, WEBM) in the box. The AI transcribes the first seconds for free, so you can check the language and the quality. If it looks right, get the full transcript: every sentence with its time, to read in the page, copy, or download as a text file or as SRT and VTT subtitles.
Is this speech to text free?
The preview is free, with no account: the first lines of your recording, written by the AI. The full transcript of the recording (text, SRT and VTT) costs €1.99, paid once for that file, up to one hour long. There is no subscription, no credit and no minutes to buy in advance.
Which AI does it use, and how accurate is it?
Whisper large-v3, the speech recognition model of OpenAI, trained on millions of hours of speech in 99 languages. On a clear recording it gets almost every word right, with punctuation and capitals. Strong background noise, people talking over each other, heavy accents and rare names lower the accuracy: the preview shows you what to expect before you pay.
Which languages can it transcribe?
99 languages, among them English, Spanish, French, German, Italian, Portuguese, Dutch, Chinese, Japanese, Korean, Arabic, Hindi, Russian and Turkish. The language is detected by itself; if the detection is wrong (a short or noisy start), choose the language in the list and the preview is made again.
Can I make subtitles for a video?
Yes. Drop the video itself: its sound is read in your browser, and only the sound is sent to the AI. The SRT and VTT files cut the speech into short subtitles of two lines at most, timed on the voice, ready for YouTube, Vimeo, VLC, Premiere Pro, DaVinci Resolve or CapCut.
Is my recording uploaded?
Only its sound, and only to transcribe it: your browser turns it into a light 16 kHz mono file and sends it in parts of two minutes to our AI provider, Replicate, which deletes them automatically within an hour. The recording is never kept, published or used to train models. The video itself never leaves your device.
Related tools
- Audio cutter: keep only the part of the recording you need before transcribing it.
- Word counter: count the words and the reading time of your transcript.