Tibetan transcription: test a short clip first
Tibetan is one of the weakest languages in the speech model, which had very little Tibetan audio to learn from. Expect a rough draft with many errors, and on some recordings the output may be mostly wrong or repeat itself. Upload a short clip first to see what you get before you send a long teaching.
Results are best with slow, clear speech from one speaker close to the microphone. Chanting, prayers recited in unison, and a teacher with a translator talking in turns are much harder. Where the draft is close, correcting it beside the player is still faster than typing Tibetan from nothing.
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.
Tibetan script in the transcript
The transcript is in Tibetan script (Uchen), with syllables separated by the tsheg. Check stacked letters, prefixes and suffixes that are written but not pronounced, and the shad at the ends of phrases, which may be missing or placed differently from a printed text.
There is no built-in Wylie or phonetic transliteration. Correct the Tibetan first, then convert it with a separate tool if you need Wylie.
Dialects and mixed languages
Spoken Tibetan varies a lot between Ü-Tsang (including Lhasa), Kham and Amdo, and the varieties can be hard for speakers of the others to follow. The single Tibetan option will not handle them equally; Kham and Amdo speech may come out weaker. Exile communities often mix in English, Hindi, Nepali or Chinese words.
For a teaching given in Tibetan with an English translator, consider transcribing the file with English selected to get the translation in text, and correcting the Tibetan separately.
What people record in Tibetan
Uploads include Dharma teachings, oral-history interviews with elders in Dharamsala, Kathmandu and communities in Europe and North America, language lessons, and family voice messages. Researchers and translators use the draft as a rough guide to locate passages in a long recording.
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 recorded teaching or a YouTube interview, see the Tibetan video to text page.