bytedance/deer-flow
- Source
- GitHub
- First trending
- Category
- Agents
- GitHub stars
- 82,552
- Main language
- Python
This page introduces an external open-source repository. It is not an HDATF product.

What it does
An open-source agent harness that researches, writes code and creates content over long tasks. It orchestrates sub-agents, memory, sandboxes, tools and extensible skills, plus a message gateway, for work that can take minutes to hours.
How it helps ATF
Worth comparing closely with Harness, which also keeps state and memory for an AI task and hands work to other executors. Its research focus could also be a reference for LabChin.
License
MIT Permissive. Commercial use and changes are allowed if the copyright notice is kept.
More in this category
- ruvnet/ruflo
An agent harness around Claude Code and Codex that adds specialized agents, coordinated swarms, memory across sessions, federated communication between machines and security guardrails. It is used through a CLI and an MCP server. - HKUDS/ClawWork
ClawWork is a benchmark in which AI agents complete professional tasks from the GDPVal dataset, pay for their own token usage and must stay solvent. A leaderboard compares agents by income, cost and average work quality. - RightNow-AI/openfang
An open-source agent operating system written in Rust and shipped as one binary. Its Hands are prebuilt packages that run agents on schedules without prompting, for tasks such as research, knowledge graph building and reporting to a dashboard. - Conway-Research/automaton
Automaton is a continuously running AI agent that generates an Ethereum wallet, provisions its own API key and runs a think, act, observe loop from a genesis prompt. It stops running if it cannot pay for its compute. - moeru-ai/airi
Project AIRI is a self-hosted AI companion app for virtual characters, inspired by Neuro-sama. It supports realtime voice chat and playing Minecraft and Factorio on Web, macOS and Windows.
Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.