Show HN: Computer use on the phone itself: a 2B model that operates Android apps
Category: ai-ml
Tags: android, on-device-ml, vision-language-model, gui-automation, llama.cpp, accessibility-service
Score: 8.0/10 (Innovation: 8, Technical: 9, Documentation: 8, Utility: 7)
Opposable is an Android app that uses a 2B-parameter GUI-Owl vision-language model to automate UI interactions via accessibility services, running entirely on-device with llama.cpp. It introduces optimizations like image token pruning, KV cache reuse, and early exit based on accessibility tree matches, plus an NPU-accelerated vision encoder. This enables private, offline computer-use agents on mobile devices.
Target audience: mobile developers, AI researchers, accessibility engineers
Repository: https://getopposable.com/blog/computer-use-on-the-phone
View on Hacker News