Koyal Speech to Text Models

Open Models for Indic Languages

 Posted on October 7, 2026  |  

A week ago, we released Koyal: open speech transcription models for Hindi, Kannada, Malayalam and Telugu. The official post has all the details. This is my personal side of it, plus a small script so you can try it on your own recordings today.

Why Koyal matters to me

I’ve believed every language deserves tools built for it, and small open models are how that becomes real. You can download them, run them yourself, and adapt them without depending on anyone’s API.

What makes Koyal special to me is that it does rich transcription, not verbatim. Real speech is full of false starts and hesitations, but what you want at the end is a document-ready transcript, with punctuation and numbers written the way you’d write them. Koyal gives you that directly, so you don’t need to build a clean-up layer on top.

Here is what that looks like. I gave the Malayalam model a 24-second clip of read speech, and this is exactly what came out:

പാലത്തിന്റെ താഴെയുള്ള കുത്തനെയുള്ള ഉയരം 15 മീറ്റർ ആണ്. ഇതിന്റെ നിർമ്മാണം 2011 ഓഗസ്റ്റിൽ പൂർത്തിയാക്കിയതാണ്, എന്നാൽ ഇത് 2017 മാർച്ച് വരെ ഗതാഗതത്തിനായി തുറന്ന് കൊടുത്തിട്ടില്ല.

There are full stops and commas, and “15”, “2011” and “2017” are written as digits instead of spelled out as words. You can paste that straight into a document.

Try it yourself (no ML background needed)

I’ll use the Malayalam model, koyal-ml-120m-1.0. This 0.1 billion parameter model (about 0.5 GB) and runs comfortably on an ordinary laptop. You don’t need a GPU. On my MacBook (M3), a 2.5-minute recording takes about 15 seconds.

Step 1: Install uv (one time)

uv is a tool that takes care of Python and every library the script needs, so you don’t have to set anything up yourself. Open a terminal and paste:

macOS / Linux

curl -LsSf https://astral.sh/uv/install.sh | sh

Windows (PowerShell)

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Close and reopen the terminal afterwards.

Step 2: Get access to the model (one time)

The Koyal models are free, but you need to accept the terms on Hugging Face before you can download them.

  1. Create a free account at huggingface.co if you don’t have one.
  2. Open the model page and click “Agree and access repository”.
  3. Log in from the terminal and follow the prompts. It opens your browser, or you can paste an access token from huggingface.co/settings/tokens (the “Read” type is enough):
uvx hf auth login

Step 3: Download the script

Save transcribe.py into a folder, and open a terminal in that folder.

Step 4: Transcribe

uv run transcribe.py my_recording.mp3

That’s it. The transcript is printed on screen and also saved next to your audio as my_recording.txt.

The first run takes a few minutes, because it installs the libraries (including PyTorch, which is large) and downloads the model. Later runs start in seconds.

A few other things you can do:

uv run transcribe.py interview.m4a lecture.wav    # several files at once
uv run transcribe.py my_audio_folder/             # every audio file in a folder

Alternative: skip the download. uv can run the script straight from the gist, so you can skip Step 3 if you like:

uv run https://gist.githubusercontent.com/kavyamanohar/bd1c1852c1a8d1232367fa746973d848/raw/transcribe.py my_recording.mp3

Only do this with scripts from sources you trust, since it runs the code without showing it to you first. You can always read it on the gist page.

Most formats work out of the box: .mp3, .m4a (phone voice memos), .wav, .ogg/.opus (WhatsApp voice notes), .flac, and even the audio track of an .mp4 video. The script converts everything to the 16 kHz mono audio Koyal expects, so you don’t have to.

No Malayalam recording handy? Grab a few clips from this sample dataset.

What about long recordings?

Like most speech models, Koyal is trained on short utterances (up to about 30 seconds), so you shouldn’t feed it a one-hour meeting in one go. The script handles this for you. Any recording longer than 30 seconds is passed through Silero VAD, a tiny voice activity detector that finds where people are actually speaking. The script then:

  • cuts only at real pauses (half a second or longer), so no word is chopped in half,
  • drops long silences, which makes it faster and avoids confusing the model,
  • packs the speech into pieces of up to 30 seconds, transcribes them in batches, and joins the text back together.

I tested this on a 2.5-minute mp3 made from ten clips. Every sentence came back, with a character error rate of 2.6% against the reference transcripts.

Already using NeMo?

If you already have NVIDIA NeMo installed, you only need three lines for a short 16 kHz WAV file:

import nemo.collections.asr as nemo_asr

model = nemo_asr.models.ASRModel.from_pretrained("adalat-ai/koyal-ml-120m-1.0")
print(model.transcribe(["audio.wav"])[0].text)

Good to know

  • Offline only. This model transcribes recordings, not live speech. For streaming, see the multilingual koyal-indic-600m-1.0.
  • Clean audio helps. Read speech and dictation work best. Noisy, conversational audio is harder, and the model card shows the spread across benchmarks.
  • Malayalam only, and not code-switched. For Hindi, Kannada or Telugu, swap the model name at the top of the script for koyal-hi-120m-1.0, koyal-kn-120m-1.0 or koyal-te-120m-1.0.

Thank you

At Adalat AI (YC F26), the research team was encouraged to work on general-purpose models, not just legal-specific ones. Koyal stands on the shoulders of existing open research, and we believe in giving back to the community that made it possible. Kumarmanas Nethil, who leads ML at Adalat, shared this vision, and top leadership backed it fully. Thank you, Utkarsh Saxena and Arghya Bhattacharya. And cheers to Kush Juvekar and Dia Krishnan, who went to any length to get the research right and turned that vision into models you can actually download.

Thank you also to Sagar Desai and Arundhati Banerjee at NVIDIA, and to the Inception Program, for the support. Koyal is what came out when all of that came together.

I’m proud of this one, and grateful to everyone who made it happen. 🎉

If you try it, please tell us what works and what doesn’t. The Community tab on the model page is the best place.

Speech to Text  ASR  Open Models 
comments powered by Disqus