Beijing Tianheshui Technology Co., Ltd. is a high-tech company focused on “AI + Public Security” and “AI + Education”. Powered by artificial intelligence, we build an intelligent protection system for early-warning protection of minors and care for the elderly, delivering round-the-clock, fully intelligent safety solutions for campuses, public venues and specific industries.
Founded in 2017 and registered in Shunyi District, Beijing, the company focuses on AI visual recognition and face recognition. Responding to the national “AI+” initiative, we work on frontier AI vision and real-time big-data analytics to build an all-weather, fully intelligent, multi-scenario AI application system — from smart campuses to smart eldercare, from urban security to social governance, putting algorithmic capability to work where people are actually protected.
Deeply engaged in public safety, we serve police, judicial and social-governance scenarios with face recognition, video structuring and real-time early warning — replacing manual watch with technical prevention.
For campuses and key venues we build a closed protection loop of identity verification, risk early warning and school–family coordination, so that protection happens before incidents occur.
We deliver one-stop “hardware + software + capability” data-centre construction and operation, covering compute platforms, PB-scale storage, Class III MLPS security and data-asset governance.
From the underlying algorithm engine to the business platform, the entire chain is developed in-house, supporting billion-scale, multi-modal and cross-decade face recognition.
Built on a national research institute’s 1 : 1.4 billion high-precision face matching engine, supporting millisecond retrieval from billion-scale face databases with automatic alerts.
Going beyond the classic one-to-one comparison, it matches a whole set of faces in a single pass, with very low requirements on image sharpness and shooting angle.
Using facial landmarks and deep learning to predict how features change with age, it opens new possibilities for people missing for years and for long-cold cases.
Real-time capture and matching on live video streams, with target trajectory search, camera-point mapping and real-time alerts — turning mass footage from “review afterwards” into “detect instantly”.
An in-house fusion of AI and blockchain for data notarisation, trusted transfer and business collaboration — a traceable technology foundation for governance data.
A “three-security” framework across capture, transmission and storage, built to Class III MLPS standards, covering personal-information protection and minors’ data compliance requirements.
Centred on the “joint protection of the young and the elderly”, we extend into police and judicial work, social governance and public services, forming a replicable, operable and sustainable product matrix.
Front-end devices and real-name verification terminals are deployed at entrances of entertainment venues, hotel front desks and around campuses to automatically identify entrants, raise immediate risk alerts and notify guardians — moving protection ahead of risk.
A care network combining early warning, monitoring and health for the ageing population, so that elderly people living alone, with dementia or left behind are continuously seen — turning sudden risks into manageable daily routine.
Command and business-support systems for detention centres, police stations and other grassroots units, closing the loop of information gathering, dispatch and performance review on a single platform.
With our own data centre as the hub, we provide compute, storage and intelligent analytics for large-scale front-end video, delivering governance data capability to government and service entry points to the public.
Provides data aggregation, compute support and intelligent analytics for a 50,000-stream AI video network covering the whole prefecture, carrying two core businesses — early-warning protection for minors and eldercare — and building an integrated protection system for the young and the elderly. The project is led by the Political and Legal Affairs Commission, funded by the enterprise and operated on market terms, at zero fiscal cost.
A two-week pilot at key venues. Using the high-precision matching engine and real-name verification terminals, it automatically detected and logged the unlawful admission of minors, confirming the system’s practical effectiveness in real-world conditions.
On its first day live, the face recognition system matched several fugitives in the wanted database; with video tracking they were all apprehended within two days. Cameras deployed in crowded areas captured, matched and alerted in real time, compressing manual needle-in-a-haystack video review into instant identification and immediate location.
View the news coverage ↗
Continuous deployments at large performance venues helped apprehend several fugitives within four months, becoming a widely followed application case and bringing AI face recognition from specialised fields into public awareness.
View the news coverage ↗
Working with a well-known family-reunification platform, we tackle reunification challenges with N-to-N set matching and cross-age appearance prediction: families often lack clear recent photos and recollections can be imprecise, while our technology tolerates low image quality and can connect to missing-person databases and urban video resources — opening new possibilities for long-cold cases.
View the news coverage ↗
Verifiable technology, verifiable credentials. Backed by joint R&D relationships with national research institutes and a complete independent IP portfolio, we give customers and projects credentials they can audit and verify.
For early-warning protection of minors at key venues, we jointly developed dedicated front-end devices with the First Research Institute of the Ministry of Public Security and Huawei Technologies, and completed platform-specific adaptation based on customised hardware–software coordination — a complete loop from device selection and platform integration to engineering delivery.
Core technology, industry applications and platform ends are all developed in-house, forming a complete technology asset system.
The core team has worked on artificial intelligence since 2011, committing to the face recognition track well before the current AI wave — laying the technical and talent foundation for the company to come.
Beijing Tianheshui Technology Co., Ltd. is founded in Beijing. In the same year the company completes core products such as the AI face recognition system and makes an early move into AI–blockchain fusion, building a multi-product technology asset base.
The police face recognition system proved effective from the day it went live; large-venue deployments drew wide attention; and the company partnered with a family-reunification platform to help find missing people with AI.
11 software copyrights registered, covering four product lines: AI face recognition, police and judicial command management, AI + blockchain, and social governance & public services.
The intelligent early-warning protection system for minors at key venues completed its pilot run; its practical effectiveness was validated and it formally entered roll-out.
Joint development of dedicated devices with the First Research Institute of the Ministry of Public Security and Huawei; launch of the “joint protection of young and old” supporting data centre project in a prefecture in southwest China, building an integrated protection system for the young and the elderly.
“Your future comes because of us. Let us build a warm, intelligent future together.”