HANGZHOU, China, Sept. 24, 2026 /PRNewswire/ — Amap Platform, Amap’s global provider of location-based solutions, has launched Qianyu, a spatial-intelligence open platform that incorporates its location and mapping capabilities into standardized agentic tools. Developers and enterprises can deploy agents through natural-language prompts, with no coding required.
The announcement came at the 2026 Apsara Conference in Hangzhou. Amap Platform showcased how it is applying spatial intelligence to two-wheel mobility, AIoT devices, and overseas services.
Cui Yong, General Manager of Amap Platform, said the company had transformed spatial intelligence into ready-to-use agents and made them available to developers and enterprises, enabling AI not only to understand space and perceive changes but also to support decision-making for real-world businesses.
While large language models primarily focused on answering questions in the digital world, many real-world tasks—such as choosing store locations, routing fleets, and ensuring riders’ safety—also rely on understanding physical places, road networks, buildings, and how the city and surroundings change over time.
The Qianyu open platform is structured around the ‘Point-Line-Plane-Dynamic’ framework, integrating points of interest (POIs), roads, buildings, regions and time-series information into a machine-readable spatial system. As of August 2026, the platform includes more than 80 million POIs, over 5 million kilometers of roads and 7,500 square kilometers of three-dimensional models. It has expanded the information coverage to include addresses, travel times, building outlines, and area-level heatmaps.
Users can give natural-language instructions to the agent on Qianyu; the agent will then determine the spatial-temporal scope, gather information, verify with multiple sources, and then provide actionable and traceable recommendations. Tasks such as store location selection, neighborhood analysis, and demand heatmap generation usually take days to complete. With Qianyu, these tasks can be packaged into an agent-executable workflow, substantially improving operating efficiency for enterprise users, the company said.
Amap also extends features previously used in car navigation, such as lane-level guidance and traffic-light countdowns, to two-wheel mobility scenarios for scooters and motorcycles. Its ‘Eagle Eye Guardian’ alert system for two-wheel mobility provides beyond-line-of-sight warnings and flags high-risk situations including rear-end collision, dooring accidents and lane-cutting.
Its spatial solution is now compatible with Android, Linux, OpenHarmony and RTOS. The company also deepens its collaboration with mobility brands such as Niu Technologies and Ninebot.
In AIoT scenarios, Amap partners with brands including Xiaomi and Qwen Glasses to embed standardized spatial intelligence into wearables. Users can search for nearby places by voice and receive turn-by-turn instructions through smart glasses, or hail a ride from a smartwatch with a single voice command.
For developers and partners worldwide, Amap Platform continues to make its spatial intelligence capabilities available based on its World Map service. The solution is being built as a spatial-intelligence layer covering address records, route planning, navigation, and dispatch support. This applies to mobile apps and systems for enterprises’ operating and fleet-dispatch systems.
Amap said the World Map now includes more than 300 million continuously updated POIs and over 900 million address records, with support for Chinese, English and local-language adaptation.
Amap World Map service supports navigation for driving, cycling, walking and motorcycle for logistics platforms, on-demand delivery, local services and online travel operators. Amap is extending to international markets a set of navigation capabilities already available and tested in China, including junction views, lane-level guidance, voice prompts and turn-by-turn instructions grounded in real-world road-network attributes. These features can be integrated directly into enterprise applications and operating systems.
