Uyir Lab · Language Brain
AI that speaks the languages big models miss.
Uyir Lab makes AI agents accurate in the languages big models get wrong, the way they are actually spoken.
The gap
Billions of people speak languages AI handles poorly.
Big models learn from the written web, and most of the world does not talk the way the web is written. People speak in dialects, slang, code-switching and cultural shorthand that never made it into training data. So agents misunderstand them, answer in stiff textbook language, or simply get it wrong.
- languages spoken in the world today
- 7000+
- speakers across the 16 languages on our roadmap
- 897M+
- languages, one playbook
- 16
Dialects
The same language sounds different from one city to the next. Models usually learn only the standard form.
Slang and code-switching
Real conversations mix languages and scripts mid-sentence. Training data rarely does.
Cultural context
Kinship terms, honorifics, food, faith and humour carry meaning a literal translation misses.
Our approach
Big labs go wide. We go deep.
One language at a time, we collect real conversations from native speakers, with their consent, and turn them into an understanding of how the language is really used. Depth beats breadth when your users are the ones being misunderstood.
Language Brain
A layer any AI agent plugs into.
Language Brain sits between your agent and the people it serves. It understands dialects, slang and cultural context, so the agent you already have becomes accurate in languages it used to get wrong.
- 1
01
Listen
Native speakers share real conversations, stories and voice recordings, with explicit consent and human review.
- 2
02
Learn
We build language models and evaluation sets from that data, starting with Nila, our Tamil model trained from scratch.
- 3
03
Plug in
Your agent calls Language Brain to understand people and reply the way they actually talk.
Input
Jaffna Tamil, casual. A generic model often misreads the dialect verb.
The playbook
Then we repeat it, each language faster and cheaper than the last.
01
Community
Partner with native speakers, diaspora groups and local researchers.
02
Data
Collect real speech and text with consent, reviewed by people before training.
03
Models
Train and evaluate on how the language is really spoken, not just written.
04
Reuse
Pipelines, tokenizers and lessons carry over, so every next language costs less.
Effort per language
Illustrative: each language reuses the pipeline built for the last.
From the lab
Nila is training right now.
Nila 1B, our own model with a billion parameters, trained from scratch on 9.6 billion tokens of Tamil, English and Tanglish (25 billion tokens of training in all). These numbers come straight from the training run.
nila · live
Connecting to the training run…
Training loss
Languages
The languages big models miss.
We start where the gap is widest: languages with tens of millions of speakers and very little of the data big models learn from.
- தமிழ்Live
Tamil · 86M+ speakers
Spoken Tamil differs sharply from written Tamil, and dialects from Jaffna to Chennai barely appear online.
Visit Tamil → - සිංහලNext
Sinhala · 20M+ speakers
Written and spoken Sinhala are almost two different registers, and models trained on formal text sound stiff in conversation.
Join the waitlist → - پنجابی·ਪੰਜਾਬੀPlanned
Punjabi · 156M+ speakers
About 89M Punjabi speakers in Pakistan (2023 census) write in Shahmukhi, a Perso-Arabic script, but much of the language lives in speech, not text. Many speakers write online in Urdu or English instead.
Join the waitlist → - KiswahiliPlanned
Swahili · 97M+ speakers
Most of its speakers use it as a second language, and everyday speech mixes in English (Sheng in Nairobi), which formal web text misses.
Join the waitlist → - Harshen HausaPlanned
Hausa · 94M+ speakers
One of Africa's most spoken languages, yet with little web text, written in both Latin and Arabic (Ajami) script.
Join the waitlist → - አማርኛPlanned
Amharic · 60M+ speakers
Its own script (Ge'ez) and rich word forms get split into many pieces by tokenizers built for English.
Join the waitlist → - Basa JawaPlanned
Javanese · 68M+ speakers
Speech levels such as ngoko and krama change the words depending on who you are talking to, and many speakers write online in Indonesian instead.
Join the waitlist → - FulfuldePlanned
Fula · 39M+ speakers
Spoken across West and Central Africa and written in Latin, Arabic (Ajami) and its own Adlam script. Common Crawl's language detector does not even recognize it.
Join the waitlist → - Èdè YorùbáPlanned
Yoruba · 50M+ speakers
A tonal language where meaning rides on diacritics that most web text leaves out.
Join the waitlist → - भोजपुरीPlanned
Bhojpuri · 52M+ speakers
Mostly spoken, often written in Devanagari mixed with Hindi, and not recognized as a separate language by Common Crawl's language detector.
Join the waitlist → - Afaan OromooPlanned
Oromo · 45M+ speakers
45M+ native speakers, yet fewer than 1 in every 100,000 web pages is in Oromo.
Join the waitlist → - پښتوPlanned
Pashto · 55M+ speakers
Strong regional dialects and very little high-quality digital text.
Join the waitlist → - Asụsụ IgboPlanned
Igbo · 34M+ speakers
Tonal, with many dialects, and the little text online is often missing its tone marks.
Join the waitlist → - ភាសាខ្មែរPlanned
Khmer · 21M+ speakers
Written without spaces between words, so standard tokenizers and search struggle from the first step.
Join the waitlist → - Kreyòl ayisyenPlanned
Haitian Creole · 13M+ speakers
Spoken by nearly everyone in Haiti, while French dominates formal writing, so little text reflects how people talk.
Join the waitlist → - Runa SimiPlanned
Quechua · 7M+ speakers
A family of mostly oral varieties with little standardized text to learn from.
Join the waitlist →
Early access
Build with Language Brain.
We are opening Language Brain to a small group of teams building agents for these languages, and to speakers who want to help.