Languages that live mostly in speech
Some Indian languages have millions of speakers and almost no text online. Translation software has nothing to learn from, so it guesses. For these languages we work only with native speakers, and no machine draft is used.
Who comes to us
- Museums and archives with letters, records and recordings in a regional language
- NGOs and field researchers who need interviews and survey material translated
- Universities and scholars working on oral history and language studies
- Publishers, broadcasters and documentary makers
Languages
- Kutchi, Garo, Bhojpuri, Kashmiri, Marwari, Mizo and Meitei (Manipuri)
- Bodo, Dogri, Konkani, Maithili, Nepali, Sanskrit and Sindhi
- Modi script, used for Marathi records until the mid-1900s, read and transcribed by hand
If your language is not here, ask. We will tell you honestly whether we have the right person.
What we do
- Translation to and from English
- Transcribing and translating audio and video recordings
- Reading Modi script documents and transcribing them into Devanagari, with an English translation
- Subtitles for documentaries and oral-history footage
- A glossary of names, places and local terms, so the next batch reads the same
How it works
- Send a sample page or a few minutes of audio
- We match a native speaker and tell you how long it will take
- Work goes in batches, with doubtful words flagged for you, not guessed
- You get the files, plus the glossary
Since 2010 we have translated 56 million+ words for 500+ clients, in about 44 languages. We are happy to sign an NDA before you share anything. Read our NDA
Questions
Why not use AI for these languages?
Because it fails quietly. The output looks fluent and is often wrong, and in a rare language few readers can catch it. A native speaker can.
Can you work from handwritten or damaged documents?
Yes. Send a clear scan or photo. Where a word cannot be read, we mark it rather than fill the gap.
Do you work with clients outside India?
Yes. See working with us from abroad.