GH Repository · bytedance
deer-flow
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.
- stars
- 82,571
- 30-day movement
- +26488/day
- Related entries
- 60
- Connections
- 3
agent-appPythonsuperagenttypescriptlangmanusagentmulti-agentagenticharnessagentic-frameworklangchainpythonllmmakeai-agentsdeep-researchailanggraphagentic-workflownodejspodcast
deer-flow is an open-source long-horizon SuperAgent harness from ByteDance that researches, codes, and creates. It combines sandboxes, memories, tools, skills, subagents, and a message gateway to run tasks lasting minutes to hours.
You want a ready-made multi-agent harness for extended research, coding, and content-generation tasks rather than assembling one yourself.
Use it to
- Run long-horizon deep research tasks
- Delegate multi-step coding work to subagents
- Generate content such as podcasts
- Extend the harness with custom tools and skills
- Study a LangGraph/LangChain-based agent architecture
For Developers building long-running agentic applications
- Role
- agent-app
- Language
- Python
- Licence
- MIT
- Forks
- 11,393
- Open issues
- 509
- Last push
- 2026-09-17
- Latest release
- v2.0.0 · 2026-06-25
topicsagent-harnessmulti-agentdeep-researchlanggraphpythontypescript