Loading...
Thoughts on AI desktop automation, computer-using agents, and building the future of human-AI collaboration.
RPA replays selectors recorded at design time. AI desktop automation reads the accessibility tree and runs OCR at runtime. A look at how each one fails, what maintenance costs, and which workloads should stay on UiPath or Automation Anywhere.
Read moreNo model watches your display on its own. A screenshot reaches an AI only when software you installed captures one and sends it. Here is the full pipeline, what leaks into a capture, and which parts never leave your Mac.
Read moreThe accessibility tree is the labeled hierarchy macOS publishes for screen readers, and it gives an AI agent exact element coordinates instead of pixel guesses. A look at the token math against screenshots, the cases where the tree returns nothing, and how OpenOwl queries it through AXUIElement.
Read moreAccessibility and Screen Recording are the two TCC grants every macOS computer use agent needs, and both fail in confusing ways. Here is why the prompt keeps coming back, why your screenshots are black, and why the restart is not optional.
Read moreScreen control MCP servers split into read-only capture and full input control, and that split decides your security posture more than any feature list. A breakdown of the tiers, token costs, OCR fallbacks, and what breaks in production.
Read moreCodex CLI reads MCP servers from ~/.codex/config.toml. Here is the exact TOML block OpenOwl's installer writes, the two macOS permissions the server needs, and the errors you will hit on the way.
Read moreClaude Cowork runs an agent in Anthropic's cloud for team document work. A local MCP agent runs on your own Mac and drives the apps you already have open. Here is where each one wins on privacy, cost, and what it can actually reach.
Read moreLoop engineering is the practice of writing a loop that prompts Claude until a job finishes on its own. This guide points that loop at real desktop and community work on macOS with OpenOwl, including the reliability lessons we learned running one in production.
Read moreA Perplexity Computer alternative is any tool that gives an AI assistant control of a real desktop without buying Perplexity's device. We break down local versus cloud execution, what data actually leaves your machine, and where local setups break.
Read moreOpenOwl gives Claude Code eyes and hands on your Mac through 42 MCP tools, free at 50 tool calls a day with no credit card. A setup guide covering install, permissions, your first automation, and what the free tier actually covers.
Read moreMCP is a protocol for how models discover and call tools. Computer use is a capability where a model reads a screenshot and issues clicks. They solve different problems, and the token cost between them differs by an order of magnitude.
Read moreA practical comparison of MCP servers for desktop automation, covering how screenshot and accessibility-tree approaches differ, where Windows-MCP, mcp-desktop-automation, and OpenOwl each fit, and what the permission model actually requires.
Read moreOpenClaw runs skills and shell commands. OpenOwl runs 42 MCP tools that see the screen and query the accessibility tree. A side-by-side look at the architectures, the setup cost, and where each one breaks.
Read moreClaude Code can control your screen, but the feature ships with seven hard constraints: a machine-wide lock, no headless mode, plan gating, provider exclusions, macOS-only CLI support, fixed screenshot downscaling, and per-app approval. Here is what each one blocks and how to work around it.
Read moreAnthropic's computer use requires a Pro or Max plan and is unavailable on Team and Enterprise at any price. Here is the free route on macOS using an MCP server, including the daily limits and the failure modes you will hit.
Read moreClaude Code refuses to interact with your screen? Here is the exact MCP setup that gets it controlling your desktop, with a specific prompting pattern that works.
Read more