Enterprise Deep Research (EDR) is a multi-agent system designed for automated, steerable deep research and analytics within enterprise contexts. It addresses complex query decomposition, specialized information retrieval, and adaptive research refinement, offering significant benefits for data-driven insights and report generation. The system targets power users and researchers requiring sophisticated, automated research workflows.

How It Works

EDR employs a Master Planning Agent for adaptive query decomposition, supported by four specialized search agents (General, Academic, GitHub, LinkedIn) and a Visualization Agent. A core reflection mechanism identifies knowledge gaps, enabling adaptive research direction updates, optionally guided by human-in-the-loop steering commands for real-time refinement. It integrates an extensible MCP-based tool ecosystem, supporting tasks like NL2SQL and enterprise workflows.

Quick Start & Requirements

• Installation: Clone the repository, set up a Python 3.11+ virtual environment (pip install -r requirements.txt), and configure frontend dependencies (Node.js 20.9.0+ required, npm install, npm run build).
• Configuration: Requires API keys for Tavily search and at least one LLM provider (OpenAI, Anthropic, Google, Groq, SambaNova) via a .env file.
• Running: A full-stack application can be launched with python -m uvicorn app:app --host 0.0.0.0 --port 8000. Command-line research execution is available via python benchmarks/run_research.py.
• Resources: Links to video demos are provided. Backend API documentation is available at http://localhost:8000/docs.

Highlighted Details

• Multi-agent architecture with adaptive query decomposition and specialized search agents.
• Real-time steering commands enable continuous research refinement and human-in-the-loop guidance.
• Supports multiple LLM providers (OpenAI, Anthropic, Google, Groq, SambaNova) and models.
• Includes comprehensive benchmarking suites (DeepResearchBench, ResearchQA, DeepConsult) and the EDR-200 dataset for evaluating agentic trajectories.
• Demonstrated integrations include web applications and Slack workspaces.

Maintenance & Community

The project is presented as a research output with an associated arXiv preprint (arXiv:2510.17797). No direct community channels (e.g., Discord, Slack) or explicit maintenance status are detailed in the provided README.

Licensing & Compatibility

Licensed under the Apache 2.0 license, which permits commercial use and integration into closed-source projects.

Limitations & Caveats

The provided README does not detail specific limitations, alpha status, or known bugs. Setup requires obtaining multiple API keys and configuring both Python and Node.js environments, which may present an adoption hurdle.