Shruti Bodo
Speech to text for Bodo and English, on your device, with no server.
Try it now · in your browser · free, no login
Speak Bodo, English or both. Shruti Bodo writes it down on your device.
Shruti · checking this device
Audio never leaves your device. Shruti and its open-source engine run inside this page.
(or drop one on the transcript)
Load Shruti, then press Record and speak Bodo, English or both.
Speak Bodo, English or both. English words can stay in English (Latin script) and are underlined here; Bodo is written in Devanagari. While you speak, words appear once they have settled, a second or two behind you; when you stop, Shruti reads the whole recording once more for the final text.
Shruti Bodo · 29.7 MB · Bodo and English · 7 November 2026
Bodo and English in one 29.7 MB file. On Bodo conversations, about 27% fewer errors than IndicConformer Bodo, a model almost five times its size.
Shruti Bodo writes Bodo in Devanagari, and about half of the English words stay in English letters: आंनि फिसाया गुवाहाटीयाव फरायो दिनै hostel fee होनांगौ. On real conversations (DiarBench, held out, 112 turns), counting English words only in English letters, its word error rate is 39.9, against IndicConformer Bodo’s 55.0, the one rival we measured. It writes about half of the English words in English letters (50.4%); IndicConformer Bodo writes none. It runs offline on a five-year-old phone, about 11× faster than real time in 64 MB of memory with the native engine, and it invented no words on 100 silent clips.
| Model | File | IndicVoices valid | DiarBench, English in English | DiarBench, native |
|---|---|---|---|---|
| Shruti Bodo | 29.7 MB | 34.5 | 39.9 | 51.5 |
| IndicConformer Bodo | 137.7 MB | 28.6 | 55.0 | 46.9 |
Word error rate %, lower is better; DiarBench 112 turns, the same scorer for every model, measured 3 November 2026. IndicVoices valid is IndicVoices’ own held-out split: its speakers are kept out of training and its sentences blocked, but it is the same source as our training data, so it flatters us. DiarBench is real conversation from outside it: 112 turns from 25 conversations, 1,592 reference words, 409 of them English, in 97 turns. The English-in-English column scores against references that write English words in English (Latin script), and English words count only if written in English letters; the native column scores against references that write those English words in Devanagari. Other Bodo speech models exist that we have not measured; these results cover only the models listed here.
Where it falls short
- One rival measured. Whisper, SraVaani-1.0, Dolphin and Vosk do not list Bodo. Other Bodo models exist that we have not measured.
- The native reference shows the gap. IndicConformer Bodo is ahead on DiarBench native (46.9 vs 51.5) and on IndicVoices valid (28.6 vs 34.5). Character errors: DiarBench 28.99% vs 14.58%, IndicVoices valid 14.75% vs 8.34%.
- No independent read test. IndicVoices valid shares a source with our training data, and no independent Bodo read set was available.
- English words in Devanagari: about 102 of DiarBench’s 409 by automatic alignment (france as फ्रेन्स). DiarBench also glosses some local names in Latin, so a correct native spelling can score as an error.
- In a browser it was tested on 4 test clips read by an AI voice, not real speech, played into this page on one phone’s Chrome, with the language set. All 4 came out in Devanagari. Not yet tested with a live microphone. On the other Shruti languages where we measured it, plain English loses a little accuracy this way.
- No Bodo reader has reviewed this output yet.
- Loud noise: Shruti models sometimes repeat a phrase over and over in loud noise (measured on Shruti Hindi-family models, not on Bodo).
- Not tested on Bodo: noise, music, far-away or speaker-played audio, Linux, Windows and iOS. No punctuation; numbers come out as words.
Try Shruti Bodo Shruti Bodo on Hugging Face the model file, the full results and how each model was run
Use it in your app
Every Shruti model runs in Cactus’s open-source needle engine. Download the engine for your platform and the file for your language, and pass it a WAV file or 16 kHz audio.
# Command line (Mac, Linux, Windows)
# Engine: huggingface.co/Cactus-Compute/needle3, your platform
# Model: shruti-bodo.cact, from Shruti Bodo's Hugging Face repo
./needle --model shruti-bodo.cact --audio clip.wav
# for speech that is only English: --audio-language en
# Python
pip install cactus-needle
import needle
shruti = needle.Whistle(weights="shruti-bodo.cact")
print(shruti.transcribe("clip.wav")["text"])- Android, iOS and the web: call the engine’s C API.
needle_loadtakes the model file,needle_transcribea clip, andneedle_stream_transcribe_processabout a second of live audio at a time. The demo on this page makes the same calls, from WebAssembly in a Web Worker. - Leave the language unset for Hindi and Hinglish: detection does as well as setting it on Hindi, Hinglish, accented Hindi and US English. Set English when the speech is only English with an Indian accent, which detection often takes for Hindi. On a subset of Svarah: 21.13% word errors with English set, 25.75% detected.
- Keywords, one per line, favour the names, places and product words your users say.
- Cactus’s Python package (cactus-needle) sends anonymous usage counts by default; set
NEEDLE_TELEMETRY=0to turn that off. The command-line engine and the WebAssembly engine on this page make no network requests of their own.
Credits
Every Shruti model is built on Cactus Whistle and runs in the needle engine, both by Cactus Compute under Apache-2.0. The engine files on this page are Cactus’s WebAssembly build, unmodified, served with its licence and notice. Shruti Bodo’s model card on Hugging Face lists its training datasets and their licences.