DeerFlow¶

What it is¶

DeerFlow is an open-source agentic research workflow project from ByteDance focused on deep-research style information gathering and synthesis.

What problem it solves¶

It gives teams a starting point for building structured research agents instead of stitching together ad hoc search, scraping, and report-generation scripts.

Where it fits in the stack¶

Agents / Research Workflow. It sits between agent orchestration frameworks and end-user research products.

Typical use cases¶

Deep research assistants that gather and synthesize sources
Multi-step browsing and summarization workflows
Internal research copilots that need repeatable task structure

Example company use cases¶

Strategy research: compile competitor, pricing, and tooling landscape reports before major decisions.
Sales enablement: research target accounts, competitors, and public signals before outreach.
Product discovery: gather feature, documentation, and ecosystem evidence before choosing integrations.

Strengths¶

Open-source starting point from a large AI lab
Clear fit for research-oriented agent workflows
Useful reference architecture even if not adopted directly

Limitations¶

Research-agent projects often need significant adaptation before production use
Governance, caching, and citation quality still need to be designed around the core workflow

When to use it¶

When you want a reference implementation for research-heavy agents
When browsing, synthesis, and report generation are the core user workflow

When not to use it¶

When a simpler search API plus application logic is enough
When you need a stable SaaS product rather than an open-source starting point

Selection comments¶

DeerFlow is strongest when the work looks like "collect evidence, synthesize it, and produce an informed artifact."
It is not the default choice for simple CRUD automations or fast transactional workflows.
Pair it with Tavily for search, mem0 for longitudinal memory, and Browser Use for interactive browsing gaps.

Licensing and cost¶

Open Source: Yes
Cost: Free to inspect and adapt; runtime costs depend on models and infrastructure
Self-hostable: Yes

Getting started¶

Installation¶

The recommended way to start DeerFlow is via Docker:

git clone https://github.com/bytedance/deer-flow.git
cd deer-flow
make config
make docker-init
make docker-start

Usage¶

Access the DeerFlow UI at http://localhost:2026 to start creating research tasks and managing sub-agents.

CLI examples¶

# Initialize configuration and generate config.yaml
make config

# Start the application in development mode
make dev

# Run a specific research task via the CLI harness
python3 -m deerflow.harness run --task "competitor analysis for AI agents"

API examples¶

import requests

# Example of submitting a research task to the DeerFlow API
url = "http://localhost:2026/api/v1/tasks"
payload = {
    "title": "Agentic Framework Research",
    "prompt": "Analyze the top 5 open-source agent frameworks in 2026.",
    "model_config": {"model": "gpt-5-responses"}
}
headers = {"Content-Type": "application/json"}

response = requests.post(url, json=payload, headers=headers)
print(response.json())

Sources / References¶

Contribution Metadata¶

Last reviewed: 2026-05-21
Confidence: high