Free audio to text converter
Transcribe audio to textin 99 languages, free
Upload an audio or video file, or record your voice, and get an accurate transcript with timestamps in minutes. Translate it into another language, fix any word and download it as PDF, Word, TXT or subtitles. No account, no limits on files.
- No sign-up
- 99 languages
- Translation included
- PDF · Word · SRT
1 · Choose a model
Whisper listens to the first seconds and finds the language itself.
3 · Output
4 · Your audio
Drop an audio or video file here
MP3, WAV, M4A, AAC, OGG, FLAC, MP4, MOV, WEBM… up to 60 min and 500 MB
Your file is deleted as soon as its audio is read; the text is kept for 1 hour, then deleted.
Why Escalium
Everything you needto turn speech into text
On your device or our servers
Transcribe on your phone or computer with no wait, or on our servers with Whisper Large V3 Turbo for the highest accuracy. Pick per file.
99 languages
From English and Spanish to Arabic, Hindi, Japanese or Swahili, detected automatically or chosen by you.
Translation
Get the text in the language it was spoken, or translated into 40+ languages, line by line with the same timestamps.
Timestamps
Every line keeps its start time, ready to jump to a quote or to use as subtitles.
Edit before export
Correct names and words right on the page, then download the corrected version.
Every format
PDF, Word (.docx), plain text, Markdown, JSON and SRT/VTT subtitles for YouTube, TikTok or Premiere.
How it works
Audio to textin three steps
No software to install and nothing to configure: the transcription runs on your own device or on Escalium's servers, with OpenAI's open Whisper models.
- 1
Upload or record
Choose an audio or video file from your phone or computer, or record your voice right on the page.
- 2
Pick model and language
Fast or Precise, the spoken language (or automatic detection) and, if you want, a language to translate into.
- 3
Edit and download
Read the transcript, fix any word and download it as PDF, Word, TXT, Markdown, JSON or SRT/VTT subtitles.
Use cases
One transcription tool,many jobs
Podcasts & YouTube
Turn episodes into show notes, blog posts and captions that bring search traffic.
Interviews & research
Transcribe interviews and focus groups and quote them with exact timestamps.
Meetings & classes
Keep searchable notes of calls, lectures and webinars without typing.
Songs & lyrics
Get the lyrics of a song as text or as synced subtitles for a lyric video.
Voice notes
Turn WhatsApp audios and voice memos into text you can read, copy and share.
Subtitles
Export SRT or VTT and upload them to YouTube, Vimeo, TikTok or your video editor.
Languages & dialects
99 languages, many accents and dialects
Whisper was trained on 680,000 hours of speech in 99 languages, so it understands regional accents of the same language — Spanish from Spain, Mexico or Argentina, Brazilian or European Portuguese, British, American or Indian English. Dialects without their own written standard are written in the closest standard language: Moroccan Darija comes out as Arabic in Arabic script (partly normalised towards Modern Standard Arabic), and Neapolitan, Sicilian or other Italian dialects come out as Italian. Choose the main language rather than automatic detection for the best result with dialects.
Supported files
MP3, WAV, M4A, AAC, OGG, OPUS, FLAC, WEBM, MP4, MOV, M4V and most other audio and video formats, up to 60 minutes and 500 MB each.
- English
- Spanish
- French
- German
- Italian
- Portuguese
- Arabic (incl. Darija, Egyptian…)
- Catalan / Valencian
- Dutch
- Russian
- Turkish
- Chinese (Mandarin)
- Cantonese (Precise only)
- Japanese
- Korean
- Hindi
- Afrikaans
- Albanian
- Amharic
- Armenian
- Assamese
- Azerbaijani
- Bashkir
- Basque
- Belarusian
- Bengali
- Bosnian
- Breton
- Bulgarian
- Burmese
- Croatian
- Czech
- Danish
- Estonian
- Faroese
- Finnish
- Galician
- Georgian
- Greek
- Gujarati
- Haitian Creole
- Hausa
- Hawaiian
- Hebrew
- Hungarian
- Icelandic
- Indonesian
- Javanese
- Kannada
- Kazakh
- Khmer
- Lao
- Latin
- Latvian
- Lingala
- Lithuanian
- Luxembourgish
- Macedonian
- Malagasy
- Malay
- Malayalam
- Maltese
- Maori
- Marathi
- Mongolian
- Nepali
- Norwegian
- Norwegian Nynorsk
- Occitan
- Pashto
- Persian
- Polish
- Punjabi
- Romanian
- Sanskrit
- Serbian
- Shona
- Sindhi
- Sinhala
- Slovak
- Slovenian
- Somali
- Sundanese
- Swahili
- Swedish
- Tagalog
- Tajik
- Tamil
- Tatar
- Telugu
- Thai
- Tibetan
- Turkmen
- Ukrainian
- Urdu
- Uzbek
- Vietnamese
- Welsh
- Yiddish
- Yoruba
Fast or Precise?
Choose the rightWhisper model
| Fast | Precise | |
|---|---|---|
| Model | Whisper Base | Whisper Large V3 Turbo |
| Best for | Clear speech, voice notes, meetings | Interviews, podcasts, accents, music |
| Accuracy | Good | Excellent |
| Speed | Very fast | Slower (larger model) |
| Languages | 99 | 99 + Cantonese |
FAQ
Questions abouttranscribing audio
Yes. Transcription and translation are free, with no account and no credit card. Each file can be up to 60 minutes long.
With clear audio, Whisper Large V3 Turbo (the Precise model) reaches professional-level accuracy in most major languages. Background noise, music, overlapping voices and strong accents lower it; the Precise model handles them much better than Fast.
99 languages, from English, Spanish, French or Arabic to Hindi, Japanese or Swahili. Regional accents are understood well. Spoken dialects without their own spelling, like Moroccan Darija or Neapolitan, are written in the closest standard language (Arabic, Italian).
Yes. Choose "Translate to…" and a target language: the audio is transcribed and every line is translated, keeping its timestamp. If you choose the spoken language instead, the text always stays in the language that was spoken.
It is only used to transcribe it. On this device, it is never uploaded. On our servers, the file is deleted as soon as its audio is read, and the text is deleted one hour after it's ready. It is never used to train models.
PDF, Word (.docx), plain text (.txt), Markdown, JSON and subtitles in SRT and WebVTT, with or without timestamps.
Yes. On a phone the transcription runs on the phone itself with the lightest model, so there's no wait: keep the screen on until it's done. For songs, noisy audio or long files, choose Precise and it runs on our servers; then you can leave the tab and come back for the result.
Usually a fraction of the audio's length with Fast and around its length with Precise. When many people are transcribing you'll see your place in the line and an estimate.
Making videos with your transcript? Add synced subtitles with the Advanced Video Editor →