MindSearch is an open-source multi-agent framework for web search powered by LLMs, with configurable search backends and a FastAPI-based API. It provides examples for backend and frontend setup, plus multiple deployment options.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
MindSearch is described as an open-source AI Search Engine Framework with Perplexity.ai Pro performance. It aims to mimic human-like multi-agent search behavior to retrieve information from the web.
How it works
The project organizes into agents that perform web search tasks using configurable search engines. It supports different model formats (e.g., internlm_server, gpt4) and multiple web search APIs (GoogleSearch, DuckDuckGoSearch, BraveSearch, BingSearch, TencentSearch). It includes a MindSearch API powered by FastAPI and an interface for frontend interactions (React, Gradio, Streamlit).
Getting started
git clone https://github.com/InternLM/MindSearch
cd MindSearch
pip install -r requirements.txt
mv .env.example .env
# Open .env and add your keys and model configurations
python -m mindsearch.app --lang en --model_format internlm_server --search_engine DuckDuckGoSearch --asy
# Frontend options
# React proxy setup snippet as provided in README
HOST="127.0.0.1" # modify as you need
PORT=8002
sed -i -r "s/target:\s*\"\"/target: \"${HOST}:${PORT}\"/" frontend/React/vite.config.ts
# React setup (from README)
cd frontend/React
npm install
npm start
# Gradio
python frontend/mindsearch_gradio.py
# Streamlit
streamlit run frontend/mindsearch_streamlit.py
Recent releases
- v0.1.0 v0.1.0 (2024-11-05): An open-source AI Search Engine Framework with Perplexity.ai Pro performance. Technical Report: https://arxiv.org/abs/2407.20183
Changelog (2024/11/05): MindSearch deployed on Puyu with refactored agent module based on Lagent v0.5 and improved UI for multi-query search.
Traction
6912 stars, 690 forks, 57 open issues.
Behind the repo
No explicit