II-Researcher is an open-source framework for building intelligent web search and research agents. It targets developers and researchers needing to automate deep web exploration, data extraction, and comprehensive answer generation, offering multi-step reasoning and configurable LLM integrations.

How It Works

The framework employs a modular design, allowing users to select preferred search providers (Tavily, SerpAPI) and web scraping tools (Firecrawl, Browser, BS4). It supports multi-step reasoning and reflection, enabling complex research workflows. Asynchronous operations enhance performance, and configurable LLM models (via LiteLLM) allow for task-specific optimization, including advanced context compression.

Quick Start & Requirements

• Install: pip install ii-researcher or install from source.
• Prerequisites: Python 3.7+, Docker, Node.js/npm. Requires API keys for services like OpenAI, Tavily, SerpAPI, and Firecrawl.
• Setup: Environment variables must be configured for API keys and model choices. Running a local LiteLLM server is recommended for LLM integration.
• Docs: Blog Post (link is to README itself), Demo, MCP.

Highlighted Details

• Integrates with multiple search and scraping providers.
• Supports multi-step reasoning and reflection for complex queries.
• Leverages LiteLLM for flexible LLM model integration and management.
• Offers both CLI and a web interface (via FastAPI backend and Next.js frontend).

Maintenance & Community

The project acknowledges contributions from the open-source community, specifically mentioning LiteLLM, node-DeepResearch, gpt-researcher, and baml.

Licensing & Compatibility

The README does not explicitly state a license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The "legacy pipeline mode" is deprecated. The project relies heavily on external API keys, which may incur costs. Detailed performance benchmarks or specific LLM compatibility lists are not provided in the README.