> ## Documentation Index
> Fetch the complete documentation index at: https://doc.starrise.tech/llms.txt
> Use this file to discover all available pages before exploring further.

# Deep Research Web UI

> Integrate Starrise AI models with Deep Research Web UI — an AI-powered research assistant.

## 1. Introduction

Deep Research Web UI is an AI-powered research assistant that combines search engines, web scraping, and large language models to let the AI autonomously search for information, continuously dig deeper, and ultimately produce a research report.

Key features of this project:

* **Low Cost** — Achieve "deep research" results similar to ChatGPT, Perplexity, and similar products at a very low cost
* **Privacy & Security** — All configuration and API requests are handled client-side in the browser, and it can be self-hosted
* **Real-time Feedback** — Streams AI responses and displays them in the UI in real time
* **Search Visualization** — Displays the research process using a tree structure
* **PDF Export** — Export the final research report in Markdown and PDF formats
* **Multi-model Support** — Compatible with more LLM providers

Project open-source repository: [GitHub](https://github.com)

## 2. How to Use

Open [Deep Research Web UI](https://deep-research-web-ui.vercel.app/) and click the "⚙️" button in the top-right corner to open the settings dialog.

### 2.1 Configure Starrise AI LLM Service

* Register or log in at the [Starrise AI Console](https://ai.alad.com/).
* Go to **API Keys**, generate a new API key, and copy it.
* Back on the Deep Research page, in the **AI Service** section of settings, select **OpenAI Compatible**.
* Set the **Base URL** to `https://ai.alad.com/v1`.
* Paste the API key you just generated in the **API Key** field.
* In the **Model Name** field, select the model you want to use (e.g. `gpt-4o`, `claude-opus-4-7`).
* (Optional) Set the context length: if you plan to do large-scale research, it is recommended to set this no higher than the maximum context length of the selected model.

### 2.2 Configure the Web Search Model

Currently supports Tavily and Firecrawl. Tavily is recommended as it provides 1,000 free searches per month:

* Register an account at [Tavily](https://tavily.com/), create a new API key in the dashboard, and copy it.
* Back on the Deep Research page, in the "Web Search Service" section of settings, select **Tavily** and paste the API key.
* (Optional) Set the language to use for searches.

## 3. Start Using

Once configuration is complete, you're ready to go! Enter a research topic, and the AI will automatically search and dig deeper, ultimately generating a research report. Use it to find primary sources, learn about topics of interest, look up news and summarize it, and more.