RD-Agent vs UFO
Which should you use? A side-by-side look at what each one does, who it suits, what it costs, how hard it is to set up, and where each falls short.
Open source · MIT
An AI agent framework that automates parts of data science and R&D work, such as ML competitions and quant res
Open source · MIT
Open-source Windows agent that uses an LLM to operate desktop apps, and a multi-device orchestration layer cal
The short answer
Choose RD-Agent if you need…
- Researchers who want to test automated ML engineering agents on benchmark-style tasks
- Data scientists exploring agent-driven model and feature development
- Quant researchers interested in automated factor and strategy research
- Teams comparing LLM backends for autonomous experimentation
Choose UFO if you need…
- Developers experimenting with LLM-driven Windows desktop automation
- Researchers studying multi-agent task orchestration
- People who want to automate repetitive tasks in Windows apps and are comfortable with setup
- Teams exploring cross-device workflows, using the Galaxy framework
Side-by-side
| RD-Agent | UFO | |
|---|---|---|
| Type | AI model / agent | AI model / agent |
| Price | Free and open source | Free and open source |
| Licence | MIT | MIT |
| GitHub stars | 15.2k | 10.2k |
| Setup difficulty | Advanced | Advanced |
| How you get it | Install with a package manager | Clone and build from source |
| Main language | Python | Python |
| Last code update | 15 days ago | 6 days ago |
What each one is
RD-Agent
R&D-Agent is a Microsoft open-source framework that uses large language model agents to automate research and development tasks, with a focus on data and models. The README describes it as letting AI drive data-driven AI work. The README also shows it applied to machine learning engineering benchmarks (MLE-bench) and to quantitative trading strategy research (RD-Agent(Q)).
UFO
UFO³ is a Microsoft research project for building AI agents that control Windows computers and coordinate work across several devices. Its older part, UFO², runs on a single Windows machine and uses a language model to read the screen and act on apps through clicks and API calls. The newer Galaxy part splits a request into a graph of tasks and assigns them to different devices, such as Windows, Linux, or Android, as described in the README.
Best for
RD-Agent
- Researchers who want to test automated ML engineering agents on benchmark-style tasks
- Data scientists exploring agent-driven model and feature development
- Quant researchers interested in automated factor and strategy research
- Teams comparing LLM backends for autonomous experimentation
UFO
- Developers experimenting with LLM-driven Windows desktop automation
- Researchers studying multi-agent task orchestration
- People who want to automate repetitive tasks in Windows apps and are comfortable with setup
- Teams exploring cross-device workflows, using the Galaxy framework
Strengths
RD-Agent
- Backed by Microsoft Research with linked technical report and peer-reviewed papers cited in the README
- Documented benchmark results on MLE-bench, with run details published online
- Supports multiple LLM providers through LiteLLM, per the README
UFO
- Open source under the MIT license with a public paper and documentation site
- Two tiers: UFO² is described as simpler and stable, while Galaxy adds multi-device orchestration
- Combines GUI actions with API calls, which the README says can reduce LLM calls
Not ideal for
RD-Agent
- People wanting a simple, ready-made app with no setup, since the README describes Python packages, documentation, and configuration rather than a finished end-user product
- Users who need results without paying for LLM API usage, since the agents rely on language model backends
- Non-Linux users, since the PyPI badge lists the platform as Linux
UFO
- Non-technical users wanting a ready-made consumer app, since the README describes setup and configuration steps
- Anyone needing a stable, fully tested production system, because the README labels Galaxy as in active development
- Users who need macOS-only automation, since the README lists Windows, Linux, and Android as supported device types
Trade-offs
RD-Agent
- Results and costs depend on the LLM backends used, and the README reports benchmark numbers with standard deviations rather than guarantees
- Installation and configuration details were not fully visible in the truncated README, so setup may take effort
- Some scenarios are limited; the web UI currently excludes the data_science scenario
UFO
- Depends on a language model and its API access, which the README discusses but this summary does not detail
- Galaxy setup is described as moderate in difficulty, and it is still in active development
- Primary focus is Windows, so cross-platform reach depends on the Galaxy device options
Getting started
RD-Agent
Install from PyPI (package named in badge)
pip install rdagentUFO
Clone the repository
git clone https://github.com/microsoft/UFO.gitFrequently asked questions
What is the difference between RD-Agent and UFO?
RD-Agent: An AI agent framework that automates parts of data science and R&D work, such as ML competitions and quant res. UFO: Open-source Windows agent that uses an LLM to operate desktop apps, and a multi-device orchestration layer cal. RD-Agent suits researchers who want to test automated ml engineering agents on benchmark-style tasks, while UFO suits developers experimenting with llm-driven windows desktop automation.
Are RD-Agent and UFO free?
RD-Agent: Free and open source. UFO: Free and open source.
Is RD-Agent or UFO more popular?
RD-Agent has more GitHub stars (15.2k) than UFO (10.2k). Popularity isn't the same as fit — check which one matches your use case above.
What are other alternatives to RD-Agent and UFO?
Other options people compare them with include Open Deep Research, DeerFlow, SWE-agent, Open SWE, Self-Operating Computer.
Keep comparing
Built from our catalogue entries for both products: open-source details come from each project's README (AI-assisted, refreshed weekly) and GitHub data; AI-tool details are editorial. Spotted something wrong? Let us know from either product's page.