Introduction

Crazy-Deep-Research-GPT is a hyper-functional, sleek, and fully local web-based AI interface built on Node.js and powered by Ollama. Designed for deep exploration or quick AI responses, it supports multiple intelligent research modes alongside customizable AI behaviors. This open-source project utilizes HTML-to-Markdown conversion and emphasizes user privacy with zero tracking or API keys required.

Key Features

• HTML-to-Markdown Conversion: Uses Turndown via Cheerio for seamless transformation of web content to Markdown format, enabling readable and exportable AI outputs.
• Ollama Model Integration: Supports local large language models like qwen2.5vl:3b and deepseek-r1:1.5b for powerful AI capabilities without requiring cloud-based APIs.
• Multiple Research Modes: Includes Normal, Deep Think, Search, Deep Research Lite, Heavy Duty Deep Research, and Auto modes for tailored responses ranging from quick answers to comprehensive research.
• Customizable AI Settings: Users can tweak system prompts, select primary and auxiliary models, adjust temperature for creativity, and manage WebSocket concurrency limits.
• Fully Local and Private: All AI processing is local, ensuring no tracking and no need for API keys, preserving user data privacy.
• Responsive UI: Built with Tailwind CSS, the interface is polished and mobile-friendly, ensuring smooth user experience.
• Location-Based Support: Enables AI queries involving local context such as nearby restaurants or weather updates.

Installation Guide

To install Crazy-Deep-Research-GPT on your machine, follow these steps:

• Clone the Repository:

  git clone https://github.com/crazystuffxyz/Crazy-Deep-Research-GPT.git
  cd Crazy-Deep-Research-GPT

• Install Dependencies:

  npm install

• Configure Ollama:
  Ensure Ollama is installed and running on your machine. Pull the recommended models:

ollama pull qwen2.5vl:3b
ollama pull deepseek-r1:1.5b

Verify these models appear in the advanced settings of the interface.

How to Use

After installation, start the local server and interact with your personal AI:

• Run the server with npm start.
• Open your web browser and navigate to http://localhost:3000.
• Select one of the intelligent AI modes such as Normal, Deep Think, Search, Deep Research Lite, or Heavy Duty Deep Research.
• Type your query in the input area and send it to receive AI-generated responses in the chat window.

Code Examples

const express = require('express');
const app = express();
app.use(express.json());

app.post('/api/query', async (req, res) => {
 const userQuery = req.body.query;
 const response = await runOllamaModel(userQuery);
 res.json({ answer: response });
});

app.listen(3000, () => {
 console.log('Server running on http://localhost:3000');
});

This snippet illustrates a basic Node.js Express server setup for handling AI queries using Ollama models locally.

Contribution Guide

Contributions to Crazy-Deep-Research-GPT are welcome, especially for improving AI integration, UI/UX polishing, and extending feature sets. Contributors should adhere to the MIT License guidelines and maintain respect for the privacy-first architecture. Ensure pull requests include clear descriptions and tests where applicable.

Community & Support

Community support is primarily managed through the GitHub repository’s issues section. Users and contributors can report bugs, request features, and discuss improvements. The open-source nature encourages collaboration and knowledge sharing.

Conclusion

Crazy-Deep-Research-GPT is a privacy-centric, locally run AI research interface ideal for developers and enthusiasts seeking powerful AI-assisted research without reliance on online APIs. Its flexibility, mode variety, and customization make it a standout open-source tool.

Resources

• GitHub Repository: Crazy-Deep-Research-GPT