This project provides an AI-powered research assistant that performs deep, iterative research on any topic, generating comprehensive markdown reports. It's designed for researchers, students, and anyone needing in-depth information synthesis, offering a structured approach to exploring complex subjects.

How It Works

The system iteratively refines research by generating targeted search queries based on user input, breadth, and depth parameters. It evaluates source reliability using a scoring system (0-1) and prioritizes high-scoring sources. Less reliable information is verified, and the process can generate follow-up questions to clarify research needs, creating a recursive loop for deeper exploration until the desired depth is reached.

Quick Start & Requirements

• Install: git clone https://github.com/Ozamatash/deep-research && cd deep-research && npm install
• Run CLI: npm run start
• Run HTTP Server: npm run start:http
• Dependencies: Node.js, API keys for OpenAI, Anthropic, Google, or xAI (configured in .env.local).
• Integration: Can be added to Claude Desktop via the Model Context Protocol (MCP).
• Local Firecrawl: Option to use a local Firecrawl instance with searXNG for search.

Highlighted Details

• Iterative research with configurable depth and breadth.
• Source reliability scoring and prioritization.
• Generates detailed markdown reports with findings and sources.
• Supports multiple AI model providers (OpenAI, Anthropic, Google, xAI).
• Available as an MCP tool for AI agents.

Maintenance & Community

No specific community channels or notable contributors are mentioned in the README.

Licensing & Compatibility

• License: MIT License.
• Compatibility: Permissive MIT license allows for commercial use and integration into closed-source projects.

Limitations & Caveats

The MCP version currently does not ask follow-up questions. The effectiveness of source reliability scoring and the quality of generated reports are dependent on the underlying AI models and search results.