Open-source super agent harness from ByteDance orchestrating sub-agents, memory and sandboxes with extensible skills, aimed at long-horizon tasks running hours or days. Has about 74,000 stars.
ByteDance's open-source super agent harness that runs sub-agents, memory and sandboxes for long-horizon tasks.
Updated EU status checked Last shown
DeerFlow at a glance
- What is DeerFlow?
- DeerFlow is an agent from ByteDance. ByteDance's open-source super agent harness that runs sub-agents, memory and sandboxes for long-horizon tasks.
- Who makes DeerFlow?
- DeerFlow is made by ByteDance.
- Does DeerFlow work in the EU?
- Available in the EU. MIT-licensed self-hosted software that can be cloned and run anywhere, with no regional restrictions in the license or README. Model providers used with it are configured by the user. Last verified 14 Nov 2026.
- How much does DeerFlow cost?
- Open source, open source. Free MIT-licensed software to self-host; users pay for their own model/API usage. Licence: MIT.
- Where was DeerFlow shown?
- In 1 creator video by Matthew Berman. Most recently by Matthew Berman on 24 Jun 2026, at 1:20 in “You NEED to try these 12 open-source AI projects RIGHT NOW”.
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
DeerFlow (Deep Exploration and Efficient Research Flow) is an open-source agent harness from ByteDance. Version 2.0 is a ground-up rewrite that shares no code with v1, which began as a Deep Research framework and is still maintained on the 1.x branch. It is built on LangGraph and LangChain.
It provides a runtime for agents with a filesystem, memory, extensible skills, sandbox-aware execution, tools, a message gateway, and the ability to plan and spawn sub-agents for multi-step work. Tasks can range from minutes to hours, covering research, coding and content creation such as slide decks, dashboards and data pipelines. It supports MCP servers and integrates with BytePlus's InfoQuest search and crawling toolset.
DeerFlow is self-hosted: users clone the repository, run an interactive setup (make setup), and start it via Docker or local development at localhost:2026. It needs the user's own model credentials. The README recommends Doubao-Seed-2.0-Code, DeepSeek v3.2 and Kimi 2.5, and warns to read the security notice before exposing it beyond localhost.
- Orchestrates sub-agents for multi-step, long-horizon tasks
- Sandboxed execution with filesystem and persistent memory
- Extensible skills, tools and MCP server support
- Message gateway and InfoQuest search/crawling integration
- Built on LangGraph and LangChain; Docker or local setup
As seen on YouTube
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Similar AI tools
Closest neighbours by meaning, not just by category.
DanceOPD
Research · ByteDance
ByteDance research framework that distills multiple specialist image models into one via on-policy velocity matching.
Seed 2.1
Model · ByteDance
ByteDance Seed's closed multimodal model family (Pro and Turbo) aimed at agentic, real-world productivity tasks.
Dynamic Workflows
App · Anthropic
Claude Code feature where Claude writes orchestration scripts that run tens to hundreds of parallel subagents with verification.
DeepSeek harness
App · DeepSeek
DeepSeek's desktop agentic harness for running the DeepSeek models on coding projects, reportedly several in parallel.
AutoSwarm
Agent · HKUDS
Open-source layer that coordinates multiple coding and general agents (Codex, Claude Code, etc.) in parallel with shared context.
Bernini v2
Model · ByteDance
ByteDance's open-weights video generation and editing model pairing a Qwen2.5-VL planner with a Wan2.2 renderer.
