{"type":"post","id":"agent-desktop-desktop-automation-for-ai-8490a6","slug":"agent-desktop-desktop-automation-for-ai-8490a6","title":"agent-desktop: Desktop Automation for AI","description":"Vercel Labs documents agent-desktop, a CLI and C library for agents to inspect and control desktop applications through operating-system accessibility trees.","retrieval_nugget":"Vercel Labs documents agent-desktop, a CLI and C library for agents to inspect and control desktop applications through operating-system accessibility trees. It provides a structured alternative to pixel-based UI automation, but only for interfaces that expose accessible elements and only within a caller-controlled automation loop.","published_at":"2026-11-02","updated_at":"2026-11-06","record_date":"2026-11-02","date_kind":"published_at","topics":["api-design"],"entities":["Vercel Labs","agent-desktop","agent-browser"],"related_patterns":["agent-ready-software-exposes-capabilities"],"novelty_label":"new_fact","what_changed":"The docs clarify agent-desktop reads accessibility trees rather than sending screenshots to a model, and works independently of agent harnesses, including workflows such as CAPTCHA solving.","prior_refs":["https://newruntime.com/posts/vercel-labs-releases-agent-desktop-1-0-an-368178/"],"source_url":"https://agent-browser.dev/desktop","source_urls":["https://agent-browser.dev/desktop"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:agent-desktop-desktop-automation-for-ai-8490a6","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/agent-desktop-desktop-automation-for-ai-8490a6/","markdown":"https://newruntime.com/posts/agent-desktop-desktop-automation-for-ai-8490a6.md","json":"https://newruntime.com/posts/agent-desktop-desktop-automation-for-ai-8490a6.json"},"markdown_url":"https://newruntime.com/posts/agent-desktop-desktop-automation-for-ai-8490a6.md","next_reads":[{"type":"topic","path":"/topics/api-design/","url":"https://newruntime.com/topics/api-design/","title":"Api Design","reason":"Explore the api-design topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/api-design.md","json_url":"https://newruntime.com/topics/api-design.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}