DeerFlow

DeerFlow (Deep Exploration and Efficient Research Flow) is an open-source "super agent harness" from ByteDance. It orchestrates sub-agents, memory and sandboxes, extended with skills, to handle long, multi-step tasks such as deep research. Version 2.0 is a ground-up rewrite; the original deep-research framework lives on its 1.x branch.

DeerFlow runs sandboxed code locally, in Docker or in Kubernetes pods, and gives agents a file system, browser control, sub-agents, session goals and context compaction. It can retrieve from private knowledge through RAGFlow, and its website shows real demo runs. It reached #1 on GitHub Trending when version 2 launched.

DeerFlow is one of the most capable open-source agent setups for long research and multi-step tasks you can run yourself. It is MIT-licensed.

Stars

0

Forks

0

Commits

0

Repository

bytedance/deer-flow

• ManusGeneral-purpose autonomous AI agent.
• ChatGPTOpenAI's AI assistant, with agent and deep research modes.
• KortixNo. 110Open-source platform for building and running AI agents.