Dev builds model-agnostic macOS-controlling AI agent in Daytona sandbox, tests 10 real tasks
jamesmurdza · x · 2026-10-28
Developer jamesmurdza demonstrated an AI agent that controls macOS inside a Daytona sandbox and works with any LLM. He tested it with Claude, Gemini, Qwen, DeepSeek and GPT across 10 real-world tasks — filing taxes, using Apple Maps, researching and writing notes, filling web forms, changing system settings, managing contacts, and adding calendar events — each with a demo video. The setup effectively makes computer-use agents model-agnostic while keeping host-machine risk sandboxed.
Related event: Dev Builds Sandbox AI Agent to Control macOS, Benchmarks 5 LLMs(2 posts)→
More from coding & agent
- MCP for Blender teams up with Meshy to bring best-in-class 3D generation into Blender workflows — sidahuj · 2026-10-28
- Nous Research sees nonstop signups after integrating Memelord's MCP — NousResearch · 2026-10-28
- Garry Tan touts GBrain's customer comms agent: drafts replies, recaps and win-backs — garrytan · 2026-10-28
- Creator Studio launches in-app AI assistants, MCP support, open-sources Keynote format — rxwei · 2026-10-28
- Stringman: open-source $795 cable robot covers a whole room, with MCP server — GregKamradt · 2026-10-28
- Proof of Agent: paste a prompt and your AI mines, grades and governs on its own — adamamcbride · 2026-10-28