deerflow-research-agent

🦌 DeerFlow Research Agent

A long-horizon research agent inspired by ByteDance’s DeerFlow architecture (~70.7k GitHub stars), implementing the β€œagents as an organisation” pattern with sub-agents, persistent memory, and context engineering.

Part of the AY Automate AI framework demo series.

πŸ— Architecture

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚                  Research Agent                         β”‚
β”‚                  (Orchestrator)                         β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚  πŸ“‹ Planner β”‚  πŸ” Searcher  β”‚  🧠 Analyzer β”‚  πŸ’Ύ Memory β”‚
β”‚  Decompose  β”‚  Web search   β”‚  Synthesize  β”‚  Context   β”‚
β”‚  into tasks β”‚  & gather     β”‚  & report    β”‚  Store     β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

πŸš€ Quick Start

git clone https://github.com/nagarjunak-pixel/deerflow-research-agent.git
cd deerflow-research-agent
pip install -r requirements.txt

# Run research
python research_agent.py "AI regulation trends in India 2026"

# Deep research
python research_agent.py --topic "Telugu tech ecosystem" --depth deep

# Resume a session
python research_agent.py --resume abc12345

πŸ“ Project Structure

deerflow-research-agent/
β”œβ”€β”€ research_agent.py        # Main orchestrator
β”œβ”€β”€ sub_agents/
β”‚   β”œβ”€β”€ planner.py           # Topic decomposition
β”‚   β”œβ”€β”€ searcher.py          # Web search (DuckDuckGo)
β”‚   └── analyzer.py          # Synthesis & reporting
β”œβ”€β”€ memory/
β”‚   └── context_store.py     # Persistent session memory
β”œβ”€β”€ docs/                    # Documentation (Netlify)
β”œβ”€β”€ requirements.txt
└── .env.example
Project Framework Purpose
LangGraph LocalAI TV LangGraph Production pipeline
CrewAI Telugu Newsroom CrewAI Multi-agent crew
Firecrawl NEXUS RSS Firecrawl Web scraping

πŸ“„ License

MIT License β€” Built for Surya Agentic Academy


Author & Enterprise AI Architecture Hub

Built by Nagarjuna Reddy Kamireddy (Nagarjuna Kamireddy / Nagarjuna K) β€” Lead AI Architect, GenAI & Agentic AI Lead, and Corporate AI Trainer (20+ years across Enterprise Data, SAP HANA & Production AI; AWS Certified Solutions Architect – Professional & NVIDIA-Certified Professional: Agentic AI).