AgentSearch is a Python framework designed to empower developers in building customizable search agents. It facilitates the integration of various Large Language Models (LLMs) with multiple search engine providers, enabling advanced Retrieval-Augmented Generation (RAG) capabilities for tasks like summarizing results and generating new queries. The framework targets developers seeking flexible and powerful search solutions.

How It Works

The core of AgentSearch lies in its modular architecture, allowing seamless connection between LLMs (from providers like SciPhi, HuggingFace, OpenAI, Anthropic) and search engines (Bing, SERP API, AgentSearch's own dataset). This integration enables RAG workflows where LLMs process search results to provide summarized answers, identify related queries, and retrieve deeper information, leveraging the strengths of both LLM reasoning and targeted search.

Quick Start & Requirements

• Install via pip: pip install agent-search
• Requires SCIPHI_API_KEY environment variable.
• Supports LLMs from SciPhi, HuggingFace, OpenAI, Anthropic, and search providers like Bing, SERP API, and AgentSearch.
• Example usage: client.get_search_rag_response(query='latest news', search_provider='bing', llm_model='SciPhi/Sensei-7B-V1')
• Official documentation and user guide are noted as "coming soon."

Highlighted Details

• Enables building custom search agent workflows with RAG.
• Supports integration with multiple LLM providers and search engines.
• Offers standalone search functionality and API endpoint integration.
• Facilitates summarizing search results and generating related queries.

Maintenance & Community

• Community support is available via Discord.
• Contact information for feedback and inquiries is provided via email.

Licensing & Compatibility

• The README does not explicitly state the license.

Limitations & Caveats

• A user guide is still under development.
• The specific license is not mentioned, which may impact commercial use or closed-source integration.