Nepali audio to text: what to expect
Nepali has much less training data in the speech model than Hindi or English. Clear speech from one or two people gives a usable draft, but it needs a careful read, and noisy phone calls or recordings from busy places need considerably more correction. Names, places and numbers are the usual errors.
Upload the recording as it is: MP3, M4A, WAV, OGG, OPUS, FLAC or AAC, a voice note saved from WhatsApp or Telegram, or a video file such as MP4 or MOV. You can also record from your microphone in the browser. Long files are split and processed in parallel, so a ten-minute recording usually takes a minute or two.
Devanagari spelling
The transcript is in Devanagari as written for Nepali. While reviewing, look out for:
- Short and long vowels (इ/ई, उ/ऊ) and the use of halanta, which the draft does not always get right.
- Chandrabindu and anusvara used differently from your spelling.
- Hindi spellings creeping in, since both languages share the script.
- English words written in Devanagari or kept in Latin letters.
Hindi, English and other languages of Nepal
Spoken Nepali, especially in Kathmandu and among younger speakers, mixes in a lot of English, and some speakers switch to Hindi. Keep Nepali selected when most of the speech is Nepali. Auto-detect can label Nepali as Hindi or Marathi, so set the language yourself.
Nepal has many other languages, including Maithili, Bhojpuri, Newar, Tamang and Tharu. They are not in the language list, and running them as Nepali gives unreliable text. This page is for Nepali speech.
What people transcribe in Nepali
Common files are interviews for research and journalism, podcasts, FM radio segments, lectures, and calls with family. Nepali workers and students in the Gulf, Malaysia, South Korea, Japan, Australia, the UK and the US use it for voice messages from home. Nepali speakers in Sikkim, Darjeeling and Bhutanese communities abroad use it as well, often for interviews with older relatives.
Exports and next steps
Export TXT or DOCX for notes and documents, PDF for a fixed copy, SRT or VTT for subtitles, and CSV with one row per line and its start time. Timestamps and speaker names can be switched on or off in each export. Speaker labels are available on paid plans or with a Minute Pack.
For video files, such as a YouTube podcast or a TikTok clip, see the Nepali video to text page.