About
Apollo is a crop intelligence startup built in Tashkent. Instead of adding sensors to the farm, it builds models that read imagery farms already have and diagnose plant disease and stress. The product has three parts. Scout gives a diagnosis from a phone photo of a leaf, with a confidence score and recommended next steps, on the web and in Telegram, in local languages. Fleet turns standard drone imagery into row-level stress maps that flag plants needing attention before symptoms are visible from the ground. Grid adds scan history, alerts and an API so that crop health data can be used in lending, insurance and procurement decisions. The core model is a ResNet-18 classifier trained on the PlantVillage dataset of more than 70,000 leaf images across 38 conditions, with a language model that turns results into plain guidance. Apollo names smallholders, agribusiness, banks and insurers, and food security programmes as its users. In 2025 it won first place at the AI500 hackathon hosted by Agrobank, and its 2026 plans include multi-crop coverage and a mobile app. Apollo took part in the second batch of the UzCombinator accelerator.