Reverse-Engineered Accessibility Tree Desktop Control Tool
aigclink · x · 2026-07-14
A developer reverse-engineered Codex's Computer Use to build pi-computer-use, a tool capable of operating macOS / Windows desktop software.
Instead of guessing button coordinates from screenshots like typical computer use tools, its core approach is ax-first: it reads the app's accessibility tree and platform accessibility APIs to directly extract UI elements like buttons, text, and controls before executing click / type / scroll commands. Vision is only used as a fallback for supported models. macOS requires a helper app with Accessibility + Screen Recording permissions, while Windows interfaces directly with the platform's accessibility API.
The author notes that this isn't meant to replace app APIs or MCP. Direct connections should always be preferred when available; computer use serves as an ecosystem gap-filler, empowering agents to operate legacy software that lacks APIs.
More from coding & agent
- Alex Townsend posts 200 open problems in numerical linear algebra for humans and AI agents — IgorCarron · 2026-09-11
- Kimi K2.8 Preview rolls out: near-K3 coding performance, 1M context for all tiers — teortaxesTex · 2026-09-11
- Looking for a classifier of software engineering task shapes to pick models per task — StewartalsopIII · 2026-09-11
- Steal this idea: prompt-to-hardware where agents assemble custom devices — paraschopra · 2026-09-11
- Model Is the Least Interesting Part: A Guide to Six Core AI Architectures from RAG to Multi-Agent — goyalshaliniuk · 2026-09-11
- Non-coder builds layered memory architecture: 20k tokens tracks a year of agent conversations — matteoianni · 2026-09-11