u14app/deep-research

Use any LLMs (Large Language Models) for Deep Research. Support SSE API and MCP server.

What it solves

Deep Research is an AI-powered tool designed to automate the creation of comprehensive research reports. It reduces the time spent on manual information gathering and synthesis by generating in-depth analysis on various topics in a few minutes.

How it works

The system follows a structured four-stage process:

  1. Topic Definition: The user provides a research topic and optional local resources.
  2. Planning: The AI asks clarifying questions, develops a research plan, and generates search queries (SERP).
  3. Information Collection: The tool retrieves data from both the internet (via search engines like Tavily or Searxng) and local knowledge bases (uploaded PDFs or Office files).
  4. Report Generation: The AI summarizes the collected materials into a final Markdown report, which can be further refined or translated.

Who it’s for

It is for researchers, analysts, and anyone needing rapid, high-quality synthesis of web and local data into a structured report.

Highlights

  • Multi-LLM Support: Compatible with Gemini, OpenAI, Anthropic, Deepseek, Grok, Mistral, and local models via Ollama.
  • Privacy-First: All research data and reports are stored locally in the browser.
  • Knowledge Graph: One-click generation of a knowledge graph for a systematic overview of the report.
  • Flexible Deployment: Supports Vercel, Cloudflare, Docker, and static page deployment.
  • Extensible API: Provides SSE and MCP (Model Context Protocol) server support for integration into other AI services.

Related

  • Project
  • Project
  • Project
  • Project
  • Project