Personal-AI-Search

• "Less is More! There is no agent framework at all; only individual simple API call methods."

Project Overview

Personal Agentic search assistant based on GPT-4o and DeepSeek-R1. This project aims to provide a powerful and flexible search assistant that leverages advanced AI models to enhance search capabilities.

Before answering user questions, conduct brief thinking to generate necessary search keywords (optionally introduce an additional search-thinking Agent, e.g. DeepSeek-R1, to deeply analyze potential search pathways and keywords before extraction); then utilize Web retrieval tools through Func Call to acquire knowledge; finally, answer user questions based on retrieved knowledge (optionally introduce an additional summarization Agent, e.g. DeepSeek-R1, to perform preliminary thinking and response generation before the Master Agent's final answer); 【Lightweight Version of AI Search】

Project Structure

Personal-AI-Search/
├── README.md
├── LICENSE
├── requirements.txt
├── app/
│   ├── api/
│   │   ├── app.py
│   │   ├── api.py
│   │   └── main.py
├── config/
│   ├── __init__.py
│   ├── prompts.py
│   └── setting.py
├── core/
│   ├── agents/
│   │   ├── __init__.py
│   │   └── base.py
│   ├── llm/
│   │   ├── __init__.py
│   │   └── llm.py
│   └── tools/
│       ├── __init__.py
│       └── web_search.py
├── utils/
│   ├── log_utils.py
│   └── utils.py
└── .gitignore

Module Introduction

• app: Contains the main application logic and API interfaces.
  • api: Provides API endpoints and routing.
    • app.py: Entry point of the application.
    • api.py: Defines API routes and handling logic.
    • main.py: Core API logic.
• config: Contains configuration files and settings.
  • init.py: Initializes the config module.
  • prompts.py: Defines various prompts and messages.
  • setting.py: Application configuration settings.
• core: Contains core functionalities and logic.
  • agents: Agent modules.
    • init.py: Initializes the agents module.
    • base.py: Base agent class.
  • llm: Language model-related logic.
    • init.py: Initializes the LLM module.
    • llm.py: Implementation of language models.
  • tools: Various utility tools.
    • init.py: Initializes the tools module.
    • web_search.py: Web search tool.
• utils: Contains utility functions and helpers.
  • log_utils.py: Logging utilities.
  • utils.py: General utility functions.
• .gitignore: List of ignored files.

Usage Instructions

Installation Steps

1. Clone the repository:

   git clone https://github.com/your-repo/Personal-AI-Search.git 
   cd Personal-AI-Search

2. Install dependencies:

   pip install -r requirements.txt

3. Configure settings:
   • Edit the config/setting.py file to configure application settings.

Running the Project

1. Start the application:

   chainlit run app/api/app.py -w --port 5000

2. Access API endpoints:
   • By default, the Web Demo runs at http://localhost:5000.