Pioneering Sovereign AI for India's Public Sector.
Headquartered in Jaipur, Rajasthan, MultiLipi is an AI-first deep-tech enterprise engineering sovereign language models, historical paleography HTR, and automated legal workflows built specifically for India's linguistic diversity and complex public governance.
Who We Are & Why We Started
MultiLipi was founded in Jaipur with a singular conviction: foreign foundational models hosted abroad cannot solve India's hardest administrative problems. Building for India requires deep understanding of non-standard handwriting, regional colloquial speech, and statutory legal precision.
Dewang
Co-Founder & Technical LeadershipLeads MultiLipi's sovereign AI architecture, speech recognition models, and institutional government collaborations. Spearheaded the winning solution for Problem Statement 2 (Law Enforcement AI) in the National BHASHINI LEAP Hackathon, and oversees low-resource Indic dialect corpus pipelines.
Kunal
Co-Founder & Systems EngineeringArchitects high-throughput Indic document OCR pipelines, cursive script vision models, and enterprise cloud infrastructure. Ensures all government deployments comply with strict air-gapping specifications, role-based access controls, and DPDP Act provisions.
Our Core Specialities
We specialize in deep-tech domains where generic foundation models fail completely: obsolete cursive scripts, unwritten dialects, statutory police filings, and distorted citizen records.
Historical Script Paleography (Kaithi HTR)
Deep vision and handwriting recognition models trained on 400,000+ annotated bounding boxes, capable of decoding 400-year-old cursive cadastral land deeds with 84% accuracy.
Regional Dialect Speech AI (ASR & TTS)
Custom acoustic models for non-scheduled mother tongues (Marwari, Mewari, Hadoti, Dhundhari, Mewati, Bagri) empowering 16.34M regional speakers without standard Hindi approximations.
Law Enforcement & Legal BNS Intelligence
Voice-to-FIR conversion platform integrated with Bharatiya Nyaya Sanhita (BNS 2023) RAG, generating legally structured drafts with statutory penal code mappings in <90 seconds.
Indic Document Intelligence (दस्तावेज़ Wizard)
Layout parsing and optical extraction for distorted scans and handwritten citizen petitions, enabling automated department routing on Rajasthan Sampark portals.
Why MultiLipi's Mission Matters
Public sector AI is not a consumer chatbot feature. In government, AI decisions govern property rights, criminal justice, and democratic access.
Preserving Vanishing Administrative Memory
Centuries of ancestral land and revenue records exist only in obsolete scripts like Kaithi. When living readership vanishes, land disputes stall courts for generations. MultiLipi makes historical administrative truth digitally machine-readable.
Dignity in Native Language Governance
India's 1.4 billion citizens cannot be forced to interact with the state through English or generic Hindi. Citizens have a fundamental democratic right to lodge complaints, access welfare schemes, and speak in their natural mother tongues.
Absolute National Data Sovereignty
Public police filings, citizen grievances, and land records must never touch foreign servers. MultiLipi engineers 100% air-gapped, zero-egress pipelines that run securely inside State Data Centers (SDCs) and on-premise hardware.
Backed by Battle-Tested Enterprise SaaS Scale
MultiLipi is not an experimental research project. In addition to our public sector initiatives, we operate a high-throughput, enterprise-scale commercial localization platform at multilipi.com :
This commercial infrastructure provides sub-120ms latency, distributed edge caching, and 99.9% uptime for all mission-critical sovereign deployments.
Verified National Achievements
Winner for Problem Statement 2 (Speech-to-Text & FIR documentation) under the Digital India BHASHINI national challenge, selected by SVP National Police Academy & NCRB.
Executed formal MoU with the Central Government as one of the Top 3 selected startups under the Digital India BHASHINI framework.
Active R&D for citizen grievance layout parsing (Rajasthan Sampark) and speech corpus collection across 6 Rajasthani dialects.
Institutional Deployments & Strategic Alignment
Collaborating with Central Ministries, State Secretariats, and National Police Bodies
Partner with MultiLipi on Sovereign AI
We work directly with central ministries, state departments, academic research labs, and system integrators.