RadarTopicsBuildersWeeklyReads
Open Source Radar
microsoft/

UFO

GitHubWebsite

UFO is a Python-based multi-device orchestration framework from Microsoft, with UFO³ Galaxy for cross-device DAG-based workflows and UFO² for Windows automation. It provides galaxy and device agent components, configuration details for OpenAI/Azure OpenAI, and start-up commands in its docs.

9.4kstars
1.0kforks
77issues
MITlicense
2024since
Star historydaily snapshots by VibeCrowd

Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).

Alternatives & relatedmatched by topic overlap
langgenius/
dify
▲ 753/wk

Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.

151k24kTypeScript
calesthio/
OpenMontage
▲ 1.9k/wk

World's first open-source, agentic video production system. 12 production pipelines, 100+ tools, 700+ agent skill and production-knowledge files. Turn your AI coding assistant into a full video production studio.

45k5.5kPython
TauricResearch/
TradingAgents
▲ 693/wk

TradingAgents: Multi-Agents LLM Financial Trading Framework

96k18kPython
bytedance/
deer-flow
▲ 1.2k/wk

An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and message gateway, it handles different levels of tasks that could take minutes to hours.

79k11kPython
hiyouga/
LlamaFactory
▲ 162/wk

Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

74k9.0kPython
shareAI-lab/
learn-claude-code
▲ 715/wk

Bash is all you need - A nano claude code–like 「agent harness」, built from 0 to 1

73k12kPython
Reviewgenerated from repository data · Aug 5, 2026

What it is

UFO³ is a multi-device orchestration framework combining Galaxy (multi-device DAG-based workflows) with UFO² (Windows desktop automation). It supports cross-device collaboration and heterogeneous platform integration, and includes a unified agent interaction protocol (AIP).

How it works

The Galaxy part decomposes user requests into a ConstellationAgent-driven Task DAG, with dynamic graph evolution and asynchronous execution across devices. The UFO² part provides deep Windows OS integration, with hybrid GUI + API actions and the ability to serve as a Galaxy device agent. Configuration sections describe how to set up LLM providers and device pools, including OpenAI and Azure OpenAI options.

Getting started

Quick start sections show installation and configuration steps for both Galaxy and UFO²:

  • Galaxy start requires:
    • "pip install -r requirements.txt"
    • Copy and edit config/galaxy/agent.yaml.template to config/galaxy/agent.yaml
    • Edit config/galaxy/devices.yaml to register devices
    • Start Galaxy with: "python -m galaxy --interactive"
  • UFO² start requires:
    • "pip install -r requirements.txt" or "pip install -r requirements.txt" (as shown for UFO²)
    • Copy and edit config/ufo/agents.yaml.template to config/ufo/agents.yaml
    • Run with: "python -m ufo --task <task_name>"

Details and links to full guides are provided in the README for Galaxy and UFO², including Windows, Linux, and mobile device setup.

Recent releases

Latest releases (as listed):

  • 3.0.7 (2026-06-12)
  • 3.0.6 (2026-06-06)
  • 3.0.5 (2026-05-15)
  • 3.0.4 (2026-05-14)
  • 3.0.3 (2026-05-14)

Traction

Repository has 9399 stars and 1045 forks, with 77 open issues.

Behind the repo

This section would describe affiliated organization or startup details if linked; none are provided beyond the repository owner (Microsoft).

Caveats

License is MIT. Created 2024-01-08. Last push 2026-07-08. Primary language is Python. Open issues count is 77.

SharePost on XLinkedIn
All trending reposRevenue-verified startups →