Language intelligence for Odiaand low-resource Indian languages
We are building the AI layer for 38 million Odia speakers — tokenizer optimization, language models, speech, and translation for India's underserved languages.
What We Do
Efficient Tokenization
Our Odia-optimized tokenizer is 3x more efficient than generic alternatives. Built for Brahmic scripts with Unicode awareness. This means lower costs and better performance on every API call.
Lekhani Model Family
ଲେଖନୀ (Lekhani)
Odia-exclusive LLMs built on open-weight base models. Named after Odia literary tradition — Pada (verse), Chhanda (meter), Kavya (poetry), Mahakavya (great epic). All released under Apache 2.0 license.
Shruti & Anuvada
ଶ୍ରୁତି (Shruti) & ଅନୁବାଦ (Anuvada)
Odia speech recognition and synthesis (Shruti = "sound" in Odia) and Odia translation (Anuvada = "translation" in Odia). Built for government, media, and enterprise use.
The Odia Language Gap in AI
38 million speakers.
~38 million people speak Odia (35M native, 2011 Census). Zero commercial AI APIs exist for them.
Native speakers (approximate)
Norwegian (5M), Dutch (25M), and Greek (13M) all have robust AI support. Odia (~38M) has none.
- •Global AI models are objectively 3x+ worse at Odia than English
- •No commercial Odia-specific AI API currently exists
- •Existing academic projects use non-commercial licenses
- •~38 million speakers (35M native) remain underserved
How Maelis Solves It
- •Odia-exclusive tokenizer with 3x better efficiency
- •Full commercial license (Apache 2.0) for enterprise use
- •SLA-backed API with dedicated support
- •Built specifically for Odisha government and business needs
Our Thesis
We are Odia specialists, not 22-language generalists. Every part of our stack is designed for one language done exceptionally well, not many languages done adequately.
We offer both commercial licenses and open licensing options to meet different customer needs. Our models are available for self-hosted deployment or via our managed API — giving enterprises flexibility in how they integrate Odia AI into their products.
Founded in 2026 and based in Dhenkanal, Odisha, we serve government, enterprise, and developer customers who need reliable, commercial-grade Odia language AI.
Why Now
IndiaAI Mission
The Indian government's national AI strategy actively prioritizes Indic language technology. Maelis is positioned to capture this momentum.
DPIIT Startup Recognition
Section 140 ITA 2025 tax holiday. India's startup ecosystem infrastructure is maturing for deep-tech language AI companies.
Zero Competition
No commercial Odia-specific AI API exists today. Academic projects use non-commercial licenses. The market is wide open.
Research & Publications
Our research spans tokenizer optimization, model distillation, and dataset construction for low-resource Indian languages. All research models are released under Apache 2.0 license.
odia-eval-benchmark
121,947 rows · 7 task families · 34 datasets · CC-BY-4.0 / MIT
Consolidated from 34 Odia datasets. Org-first publication under MaelisResearch on Hugging Face. The first unified evaluation suite for Odia.
Hugging Face: MaelisResearch/odia-eval-benchmarkEfficient Odia Tokenization for Large Language Models
InternalEmpirical evaluation of tokenizer configurations for Odia focused on Brahmic-script efficiency. Findings remain internal until compliance clearance.
Distilling Odia Generation Capability from Open-Weight Models
Research CompleteExploratory research on distillation techniques for Odia language generation.
Product Architecture
Our product stack is designed for one language done exceptionally well. Every component is optimized specifically for Odia.
Models + speech + translation · extensions for search, documents & reasoning
Lekhani
ଲେଖନୀ — The Writer
Odia Model Family
Odia-exclusive LLMs built on open-weight base models. Multiple capability tiers from lightweight edge deployment to flagship reasoning.
Shruti
ଶ୍ରୁତି — The Listener
Speech Recognition & Synthesis
Odia speech recognition and synthesis built for government, media, and enterprise use.
Anuvada
ଅନୁବାଦ — The Translator
Odia Translation
High-quality translation between Odia and other languages with cultural context.
Khoja
ଖୋଜ — The Searcher
Semantic Search
Semantic search and information retrieval specifically trained for Odia content.
Patra
ପତ୍ର — The Reader
Document Understanding
Optical character recognition and document understanding for Odia text.
Artha
ଅର୍ଥ — The Reasoner
Chain-of-thought Reasoning
Advanced reasoning capabilities for complex Odia language tasks and analysis.
Plus: Khoja (search), Patra (OCR), Artha (reasoning) — the same Odia-first stack, extended.
The Team
Saurav Mahalik
Co-Founder & Head of Engineering
Bhubaneswar, Odisha, India
CS engineer with hands-on experience across mobile development, machine learning, and full-stack web technologies. Built Android applications, ML prediction models, and conducted deep learning research on ECG signal analysis. Combines technical breadth with a user-first mindset from production support experience.