GitHub – bytedance/deer-flow: An open-source SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skills and subagents, it handles different levels of tasks that could take minutes to hours.

DeerFlow 2.0 Just Hit #1 on GitHub, And It’s More Than Hype

If you’ve been anywhere near GitHub lately, you might have seen DeerFlow sitting at the very top of Trending. On February 28th, 2026, it claimed the number one spot after launching version 2.0. That doesn’t happen by accident.

You can explore the full project here:
https://github.com/bytedance/deer-flow

So what is DeerFlow, really?

At its core, it’s an open-source super agent harness. Think of it less like a single AI tool and more like a well-equipped workshop. Inside, you’ve got memory, sandboxes, skills, tools, and even sub-agents that can plan and execute multi-step tasks. It doesn’t just answer prompts. It researches. It codes. It builds.

Originally, DeerFlow started as a “Deep Research” framework. But something interesting happened. The community pushed it far beyond that. Developers began building dashboards, automating content workflows, generating slide decks, even creating full data pipelines. The tool evolved because people stretched it.

And the team noticed.

Version 2.0 is a complete rewrite, no shared code with v1. It’s now built on LangGraph and LangChain, shipping with everything included: filesystem access, sandboxed execution, memory, and extensible skills. You can use it straight out of the box, or customize it deeply if you like getting your hands dirty.

There’s also a new integration with BytePlus InfoQuest for intelligent search and crawling, which makes research-heavy workflows much smoother. Setup is flexible too. You can run it locally, use Docker, or configure sandbox modes depending on your environment.

What makes this exciting isn’t just the tech. It’s the direction. We’re moving from isolated AI prompts to systems that can actually get work done over minutes or even hours.

And if this is what open-source agents look like today, imagine where they’ll be a year from now.