agent-desktop

Key facts: agent-desktop 1.0.2, from Vercel Labs and published on October 30, 2026, is a CLI and C library that lets an agent read and operate desktop apps through the operating system's accessibility tree, with no pixel matching. It runs on macOS, Windows and Linux, and passed 4,500 stars on GitHub within days of release. Its README warns that snapshots and screenshots can contain private desktop content, so use it with care.

Plans: Free and open source under Apache 2.0.

Best for: Developers building agents that need to operate real desktop software. Who should skip it: anyone who only needs browser or API automation.

Pros

• Operates desktop apps through the accessibility tree, with no fragile pixel matching
• A CLI and C library, so it fits into agent tooling rather than demanding a whole platform
• Runs on macOS, Windows and Linux
• Free and open source under Apache 2.0, with fast community uptake

Watch out for

• Snapshots and screenshots can contain private desktop content: handle them carefully
• A library for developers, not a finished product for non-technical users
• Desktop automation can break when apps change their accessibility layout

Best for

Developers building agents that need to operate real desktop software. Who should skip it: anyone who only needs browser or API automation, or anyone uncomfortable with agents seeing their desktop.

Pricing

Free and open source under Apache 2.0.

FAQs

What is agent-desktop? A CLI and C library from Vercel Labs that lets an agent read and operate desktop apps through the OS accessibility tree.

Does it use screenshots and pixel matching? No, it uses the accessibility tree, with no pixel matching.

Which systems does it run on? macOS, Windows and Linux.

Is there a privacy warning? Yes: its README warns that snapshots and screenshots can contain private desktop content.