Vibeleaderboard
← All Intel
Intel / post

OpenClaw 2.0 Gets Native Computer-Use via Cua Driver

Source
Cua
Date
Cua@trycua
Thread · 6 parts

1/ OpenClaw 2.0 is out - we're excited to have worked with the @openclaw team to bring a first-class computer-use experience directly into OpenClaw. Now you can put that experience to work: launch an OpenClaw desktop fleet on Cua Cloud Fleets, with native computer use built for multi-player agent workflows:

2/ This is deeper than an external MCP connection. OpenClaw owns the agent tool, node policy, pairing, session lifecycle, and action contract. The bundled cua-computer provider connects that path to the Cua Driver SDK.

3/ The request path is: OpenClaw agent → computer → computer.act / screen.snapshot → cua-computer → Cua Driver SDK → desktop session The user-facing contract stays in OpenClaw while the desktop runtime is Cua Driver.

4/ We exercised the same path inside a Cua Cloud Fleet VM: Fleet VM → OpenClaw Gateway → OpenClaw node → cua-computer → Cua Driver SDK → Ubuntu 24.04 XFCE/X11 desktop

View on X
Read the full thread on X
Key takeaways · AI-distilled
  • The request path runs OpenClaw , computer tool, computer.act or screen.snapshot, the bundled cua-computer provider, the Cua Driver SDK, then the desktop session, so the user-facing contract stays in OpenClaw while Cua Driver is the desktop runtime.
  • OpenClaw owns node policy and pairing as well as the agent tool, session lifecycle and action contract, which Cua says makes this deeper than an external connection.
  • Cua exercised the same path inside a Cloud Fleet VM: Fleet VM, OpenClaw Gateway, OpenClaw node, cua-computer, Cua Driver SDK and an Ubuntu 24.04 XFCE/X11 desktop.
Terms in this piece · Glossary
  • MCP — The Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.
  • AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
  • multi-agent — Using several AI agents on one problem — splitting work in parallel, checking each other, or filling different roles like planner and reviewer.
Why it matters

A native (not MCP-mediated) computer-use integration reduces latency and control overhead for desktop workflows running through OpenClaw.

Read the source x.com
More from Cua
Recommended reads
Comments

Checking sign-in…

Loading comments…