Mitesh Khapra

[Mitesh Khapra @MiteshKhapra]

Everything we have built at @Bodhan_AI and @ai4bharat is now live. 🎉

Open-weight speech, vision & translation models for India's languages, available on Hugging Face. The same stack, hosted at some of the lowest prices in the market.

🤗 https://t.co/0FuZf37ddP

Read more about in our blogs:

>>> https://t.co/HVgMNaUlxj

API end-points live now on Bodhan AI.

>>> https://t.co/4HePUHHbcc

Shout-out to @SarvamAI, our inference infrastructure partners for helping us serve our models reliably and affordably at scale.

Dedicated to Pushpak Bhattacharyya Sir. His love for education and Indian language technology inspired all of it. 🙏

[Mitesh Khapra @MiteshKhapra]

@debjit012 @Bodhan_AI @ai4bharat Will be available on HF on 05.09.2026

[Mitesh Khapra @MiteshKhapra]

Transcribe. Translate. Recognise. And now — Synthesise.

Introducing Indic-Speak: 4B params, 23 languages, one model. Built by @Bodhan_AI and @ai4bharat

There's a bit in this video where two speakers, Parth and Dhara, have a podcast style conversation about "Attention Is All You Need" in Hindi — generated in a single pass, no stitching two voices together after the fact. Doesn't sound like TTS (worth checking out in the video).

Also handles code-mixed speech naturally, gets PAN/Aadhaar/currency right, and reads a full Tamil audiobook chapter without losing the thread.

Live on 05.09.2026 on @Bodhan_AI

[Mitesh Khapra @MiteshKhapra]

@Cryptlesh @Bodhan_AI @ai4bharat Yes, it will be a open weights model.

[Mitesh Khapra @MiteshKhapra]

@rithulbm @Bodhan_AI @ai4bharat Will be available for public use on 05.09.2026

[Mitesh Khapra @MiteshKhapra]

And here is the full 3 minute version, with every script and every hard case.

There's a moment about a minute into this longer video that I keep coming back to. A real Malayalam cricket newspaper page, packed with columns, a points table, photos, captions, everything. Watching it get pulled apart block by block and put back together as clean, structured text is the moment this stopped feeling like a lab demo to me.

[Mitesh Khapra @MiteshKhapra]

Text (Indic-Translate). Speech (Indic-Transcribe). Now vision.

IndicOCR, the new kid on the block. Built by @ai4bharat and @Bodhan_AI

Handwritten notes, degraded archives, 9+ scripts, all in a 0.8B model that holds its own against models 4-5x its size.

Live 05.09.2026.

Short video here 👇 full version in the reply.

[Mitesh Khapra @MiteshKhapra]

No Indian idea should have to wait for translation.

Introducing Indic-Translate: built to carry meaning across India's languages.

Built by @BodhanAI & @ai4bharat.

22 languages, both directions · 32K context, whole documents at once

Live on https://t.co/qX5WB8Rkuu · 05.09.2026

[Mitesh Khapra @MiteshKhapra]

India is a voice-first nation: countless languages, scripts, and accents.

Introducing Indic-Transcribe: built to understand India as she actually speaks. Built by @BodhanAI and @AI4Bharat.

🎙️ 1.2B params · 26 languages + English

Live on https://t.co/ue8j995kFb, 05.09.2026

[Mitesh Khapra @MiteshKhapra]

Six months ago at the IITM Bodhan AI Conclave, we made a promise to build the AI stack for Indian education.

From promise to practice, the journey is taking shape.

Watch this space as we reveal more of what we’ve been building, leading up to Sep 5.

@ravi_iitm @WSAI_IITM

[Mitesh Khapra @MiteshKhapra]

@pratykumar Brilliant vision! We at AI4Bharat have been laying the foundation with datasets for Indian languages, which will help in accelerating IndicLLM efforts.

Excited to collaborate with @SarvamAI to build the data & evals infrastructure needed to make this a reality! 🇮🇳💪 #IndicLLMs