deer-flow
by bytedance
DeerFlow is an open-source framework that runs AI agents able to research, write code and create files over tasks lasting from minutes to hours. It combines sub-agents, memory, sandboxes and add-on skills so one agent can plan and carry out larger jobs.
- Stars
- 84.5k +17 a day
- Licence
- MIT
- Runs on
- A small server
- Setup
- Technical
- Language
- Python
- Age
- 538 days
What you can do with it
- Run long deep-research tasks and get written reports
- Have an agent write and run code inside a sandbox
- Add custom skills, tools and chat channels, and schedule recurring tasks
Make money with it
A freelancer could offer research reports or automated content production, using the agent to cut the time each job takes.
An idea, not financial advice. Check the licence before charging for anything built on it.
Ways a business could use it
- Automate market and competitor research reports instead of paying for research time
- Prototype internal agents that draft documents, code or content for a small team
Can you use it commercially?
Free to use, change and sell, including in paid products. Keep the copyright notice.
Not legal advice. Read the project's licence file before building a product on it.
Getting it running
Clone the repository, add your model configuration, then run it with Docker (recommended) or as a local development setup.
Follow the official instructions in the project README.
Things to check
- Needs API keys for a language model, which usually cost money per use
- Gateway admin access is equivalent to code execution, so careless deployment is a security risk
- Version 2 is a full rewrite and the project changes quickly
- Setup is complex, with many configuration options
Checked 27 Oct 2026. Summary drafted by AI from the project's README.
