Cua
Cua scales computer fleets for computer-use agents. It runs training, evaluation, and data-generation workloads in parallel on Linux, Windows, macOS, and Android machines.
Background drivers send clicks and keystrokes to a chosen desktop window without moving the system cursor. One API boots machines through local runtimes or Cua Cloud. Pre-booted machines can be claimed in milliseconds from formally verified warm pools that scale to zero when idle. Users fork from snapshots, reproduce failures, and turn agent activity into training data. Linux support covers Ubuntu containers or full VMs with browser, terminal, Python, and package installs. Windows support includes Windows 11 and Server 2025 for native apps.
The project supplies an open-source MCP and CLI driver. Integrations cover agents such as Qwen, Code Factory, and Droid Clicky. It includes a sandbox component, fleets component, and Cua Bench for evaluation. The benchmark shows the best frontier agent clears just 6 of 25 expert KiCad tasks.
Open SourceMITCLISelf-hosted
20.6kstars
1.4kforks
5features
2025since
Overview
5 featuresCua sits in PulseGate's Developer Tools category. It focuses on scaling infrastructure to train, evaluate, and generate data for computer-use AI agents across multiple operating systems. It is built as an open-source project for AI researchers and developers. The project is open source (MIT). Cua is available on the command line, and it can be self-hosted.
Cua builds and maintains Cua, and it first shipped in 2025. The project is developed in the open on GitHub with 20.6k stars and 777 commits in the last 90 days. Among its 5 catalogued features are Agent Fleets, Cross-OS Support, and Background Drivers. It exposes integrations via an MCP server.
Summary written by a language model from the project’s public pages.
- Agent Fleets
- Cross-OS Support
- Background Drivers
- Snapshot Forking
- Batch Throughput
Tags
computer-use-agentagent-infrastructureos-fleetsmcp-driver
AI capabilities
Multimodal
Inference: Hybrid
Built with & integrations
Framework
React
Hosting
Vercel
Connectors
MCP
Runs on
CLISelf-hosted
Detected from
- React
- __remixManifest in the HTML
- Vercel
- x-vercel-id header · x-vercel-cache header
Trust & compliance
License
MIT
Verified signals
HTTPSOpen SourceFree tierActive maintenance
Indexing history
1What PulseGate has recorded for this listing
Frequently asked questions about Cua
- What is Cua?
- Cua focuses on scaling infrastructure to train, evaluate, and generate data for computer-use AI agents across multiple operating systems. It is catalogued under Developer Tools on PulseGate.
- Who is Cua for?
- Cua is an open-source project built for AI researchers and developers.
- Is Cua free?
- Yes — Cua is open source under the MIT license and free to use.
- What platforms does Cua run on?
- Cua runs on the command line. It can also be self-hosted.
- Is Cua still active?
- The GitHub repository shows 777 commits in the last 90 days.
- What are alternatives to Cua?
- Similar projects tracked by PulseGate include EfficiencyNext Data Connector, System Information Widgets, and CKB.EfficiencyNext Data ConnectorSystem Information WidgetsCKB
- Who makes Cua?
- Cua is developed by Cua.
- How long has Cua been around?
- Cua first shipped in 2025.
Also in Developer Tools
Same category — not a similarity match