Project Overview

The Deep Research Web UI offers an innovative solution for conducting comprehensive research on any topic, leveraging the power of AI. As an AI-driven research assistant, it integrates search engines, web scraping, and advanced large language models to facilitate deep research iterations. This project, an enhancement of the original deep-research, incorporates various improvements and fixes for an optimized user experience.

Key Features

• Safe & Secure: All configurations, API requests, and processes occur locally in the browser, ensuring privacy and security.
• Real-time Feedback: It streams AI-generated responses, allowing for immediate reflection and interaction within the UI.
• Search Visualization: Provides a clear representation of the research process through a hierarchical tree structure, supporting queries in multiple languages.
• Export Capability: Users can export their final research reports as PDFs for easy sharing and documentation.
• Broad Model Support: Utilizes straightforward prompts rather than more complex features, ensuring compatibility with a broader range of AI models.
• Docker Support: Easily deploy the application in a user's own environment with a single command.

Supported Providers

• AI: OpenAI compatible models.
• Web Search: Utilizes Tavily, which offers a similar experience to Firecrawl with an increased free quota (1000 credits per month).

Recent Updates

The repository has continually evolved, with recent enhancements including:

• Support for reasoning models like DeepSeek R1.
• Improved compatibility with various models and enhanced error handling.
• Reduced bundle sizes for efficiency.
• Broader language support for search functionalities.
• Comprehensive Docker support with improved PDF export processes.

Live Demo

Experience the functionality directly by visiting the live demo at Deep Research Demo.

Installation and Deployment

Easily set up the Deep Research Web UI through one-click deployment options or via pre-built Docker images. Comprehensive setup instructions can be found within the repository, guiding users through both self-hosting and local development processes.