Neel AcharyDec 30:- The Uttar Pradesh Police’s newly launched AI-based policing application YAKSH, inaugurated by Chief Minister Yogi Adityanath on December 28, is powered by Staqu’s JARVIS One, a next-generation multimodal artificial intelligence platform developed to strengthen crime prevention, investigation and beat-level policing across the state.
YAKSH was unveiled during the two-day senior police officers’ conference Police Manthan 2025 in a closed-door government event, underscoring the strategic and sensitive nature of the technology. The application marks a significant leap in India’s adoption of AI-led, data-driven policing systems.
Built on Staqu’s JARVIS One Multimodal AI Engine
YAKSH is developed on JARVIS One, Staqu’s advanced AI model capable of simultaneously processing video, audio and text data within a single unified system. Unlike conventional AI tools, JARVIS One operates entirely on internal police data, ensuring data sovereignty, security and compliance.
The platform enables law enforcement officers to;
- Search FIRs and case documents using natural language
- Identify suspects through facial recognition from photographs
- Match voice samples for investigative leads
- Analyse CCTV footage, audio recordings and textual records together
- Generate AI-led insights without relying on any external datasets
This multimodal capability represents a major evolution from Staqu’s earlier JARVIS deployments, which were primarily focused on video analytics, and now extends AI coverage across crime, security and public safety operations
First-of-Its-Kind Verified Data Creation at Beat Level;
A defining feature of YAKSH is its ground-up data verification architecture. For the first time in Uttar Pradesh, criminal data is created at the police station level and physically verified by beat officers before being incorporated into the AI system.
When an accused is registered in one district and linked to another, YAKSH automatically alerts the concerned district and beat officer, triggering on-ground verification at the individual’s residence. This ensures that any AI models built on top of YAKSH operate on continuously validated, real-world data, significantly improving accuracy and trustworthiness.
Advanced Criminal Ranking, Colour Coding and Network Analysis;
YAKSH introduces a scientific, AI-driven criminal ranking system, where offenders are categorised and colour-coded based on crime severity, repetition patterns, weapons used and behavioural risk indicators. High-risk and repeat offenders are prioritised through dynamic scoring models that update as new data is verified.
The platform also supports:
- AI-powered gang and network analysis
- Automatic linkage of FIRs, associates and criminal groups
- Real-time alerts on movement and jurisdiction changes
- Identification of Top-10 criminals at police station, district and state levels
CrimeGPT: Generative AI for Policing;
At the core of YAKSH is CrimeGPT, a generative AI interface that allows officers to query crime data in simple language. Investigators can search for criminals, patterns, gangs or past incidents without navigating multiple systems or memorising technical commands, making advanced analytics accessible even at the field level.
Adoption Beyond Uttar Pradesh;
The underlying JARVIS One multimodal AI framework is also being adopted by India’s Cyber Crime Coordination Centre (I4C) for cybercrime intelligence and investigation, where audio, video and textual evidence play a critical role. This reflects the scalability and national relevance of the technology beyond physical policing.
