Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
SciToolAgent: A Knowledge Graph-Driven Scientific Agent for Multi-Tool Integration
| Date | Stars |
|---|---|
| 2026-07-31 | 415 |
| 2026-08-02 | 415 |
| 2026-08-06 | 415 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
<div align="center">
<h1 align="center"> SciToolAgent: A Knowledge Graph-Driven Scientific Agent for Multi-Tool Integration </h1>
</div>
## 📌 Contents
- [⌚️ Overview](#2)
- [📦 Project Framework](#3)
- [⚡ Getting Started](#4)
- [🔧️ Installation](#4.1)
- [🚀 Quick Start](#4.2)
- [📊 Data Preparation](#5)
- [🛠️ Custom Tools](#6)
- [📜 SciToolEval](#7)
- [📝️ Cite](#8)
## 🆕 News
- **[2024-12]**: The SciToolAgent project is now available on GitHub.
- **[2025-08]**: Our research published in Nature Computational Science.
<h2 id="2">⌚️ Overview</h2>
SciToolAgent is a powerful agent framework designed to integrate diverse scientific tools with large language models (LLMs) to address the limitations of existing systems in scientific research. By combining LLMs as Planners, Executors, and Summarizers with a comprehensive scientific tool knowledge graph (SciToolKG), SciToolAgent autonomously plans, executes, and summarizes workflows for solving complex scientific tasks across multiple domains.
<div align=center><img src="figure/overview.png" width="100%" height="100%" /></div>
### Key Features:
- **500+ Tools**: Access to a vast array of tools, including web APIs, machine learning models, Python functions, knowledge databases, and custom tools for tasks in various scientific domains.
- **SciToolKG**: A comprehensive knowledge graph that models relationships among hundreds of scientific tools from biology, chemistry, and materials science. It encodes tool dependencies, prerequisites, and compatibility, enabling informed tool selection and combination.
- **LLM-based Planner**: Utilizes SciToolKG to autonomously plan tool sequences for problem-solving.
- **LLM-based Executor**: Executes the planned tools in sequence, retrying where necessary to ensure accurate results.
- **LLM-based Summarizer**: Compiles and synthesizes outputs from multiple tools, generating a final solution while assessing the process for improvements.
- **Safety Checking**: A built-in safety system that monitors tool execution to prevent harmful outcomes and ensures responsible research.
<h2 id="3">📦 Project Framework</h2>
```plaintext
/SciToolAgent
├── app # Application directory
├── data # Data storage directory
├── KG # SciToolKG related files
├── scripts # Scripts for running SciToolAgent
├── test # Testing scripts
├── SciToolEval # SciToolEval related files
└── tools # Tool service directory
├── DataFiles # Data file storage directory (e.g., cif, csv, md, pdb, pdf, sdf, etc.)
├── LogFiles # Log files storage directory
├── TempFiles # Temporary files directory
├── TestCode # Tool test code directory
├── ToolsFuns # Core functional modules directory
├── utils # Common utility functions for tools
├── README.md # Project overview documentation
├── requirements.txt # Python dependency list
├── run.sh # One-click execution script for tool service
├── struct.md # Project structure documentation of tool service
├── config.py # Configuration file of tool service
├── example.env # Environment variable configuration file of tool service
└── tool_runner.py# Tool execution entry script
```
<h2 id="4">⚡ Getting Started</h2>
<h3 id="4.1">🔧️ Installation</h2>
1. **Clone the repository**
First, clone the project to your local machine:
```bash
git clone https://github.com/HICAI-ZJU/SciToolAgent.git
cd SciToolAgent
```
2. **Create and activate a virtual environment**
Set up a new virtual environment using Conda and activate it:
```bash
conda create -n SciToolAgent python=3.10
conda activate SciToolAgent
```
3. **Install project dependencies**
Install the necessary dependencies for the project:
```bash
pip install -r requirements.txt
```
Optional: If you only need to use the agent part of SciToolAgent and do not require Excerpt of 8,004 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:6a6f0acd286134b7, desc:knowledge graph