agent-desktop 1.2.0 Turns the Accessibility Tree into 67 MCP Tools — Overlay Latency Still Wins
On 5 November 2026 Vercel Labs shipped agent-desktop 1.2.0. The release notes are short. The practical change is that agent-desktop mcp now serves a Model Context Protocol server over stdio and exposes every leaf CLI command as a tool. Tool names and argument schemas are generated from the clap definitions at runtime. A call returns the same JSON envelope the CLI already produced. Agents that…
On 5 November 2026, Vercel Labs released agent-desktop 1.2.0. This update simplifies the accessibility tree of desktop applications on macOS, Windows, and Linux. Agent-desktop now serves as a Model Context Protocol (MCP) server over stdio, exposing every leaf CLI command as a tool. Tool names and argument schemas are generated from clap definitions at runtime, allowing clients configured with stdio MCP to access the desktop through structured, typed data.
On Windows, this means using UI Automation (UIA) to interact with applications without relying on screenshots. This release adds an MCP surface to the existing accessibility tree functionality. Clients connect using a simple stdio MCP configuration, and tools are named "desktop_" followed by the command name with hyphens replaced by underscores.
The server handles both the current stateless MCP protocol and legacy initialize handshakes, and supports stdin for type, set-value, and clipboard-set operations. On Windows, the binary carries the standard application manifest, eliminating issues with scheduled tasks or overlay processes. Overall, agent-desktop 1.2.0 aims to provide faster and more reliable access to desktop applications by leveraging the accessibility tree.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
