convez.ai
MCP ServerBusca & PesquisaAvançado
achuthprince004

Deep Research MCP

achuthprince004/deepresearch-mcp

Pesquisa profunda automatizada com planejamento, busca multi-etapa e geração de relatórios

Ver no GitHub
InicialPluginScore · de 0 a 100

Servidor MCP que capacita agentes de IA (especialmente Claude) a realizar pesquisa estruturada multi-etapa na internet, extrair conteúdo de páginas, analisar fontes com LLM e gerar relatórios em Markdown automaticamente. Integra múltiplos provedores de busca (Tavily, SerpAPI, Google, DuckDuckGo) e análise com Groq.

Instalação

$npx convez-ai achuthprince004/deepresearch-mcp

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
3
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 4 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Pesquisa automatizada para análise competitiva e inteligência de mercado
  • Coleta e síntese de informações de múltiplas fontes para relatórios técnicos
  • Decomposição de tarefas de pesquisa complexas em workflows estruturados

Veredito

MCP server robusto para pesquisa profunda que resolve o problema real de agentes executarem pesquisa web confiável e estruturada. Suporte a múltiplos provedores e análise LLM são pontos fortes; impacto limitado por cold starts em tier gratuito e exigência de Claude web ou configuração remota.

README

A tradução para o português deste README ainda está na fila.

Deep Research MCP

A powerful Model Context Protocol (MCP) server that enables AI agents to perform structured, multi-step internet research and generate comprehensive reports automatically.

Deep Research MCP equips Claude web only with advanced research capabilities including query planning, web search, content extraction, source analysis, and report generation.

Link for endpoint generation: https://deepresearch-mcp.vercel.app/

[!IMPORTANT]

Current Status

Deep Research MCP is available as a Remote MCP server and is designed to provide advanced research capabilities through the Model Context Protocol.

The server supports end-to-end research workflows, including planning, web search, content extraction, source analysis, and report generation.

The project is actively maintained and continues to evolve with new research features, improved performance, and enhanced reliability.

Since this instance is deployed on Render Free tier, the first connection or tool call after 15 minutes of inactivity may take 30–90 seconds (cold start). The same applies for the first connection in claude web connectors too.

Subsequent calls during an active session are much faster.

This delay is normal for free hosting and mainly affects Claude when starting a new research task after a pause.

Features

  • Deep multi-step internet research

  • Research planning and task decomposition

  • Multi-provider web search

    • Tavily
    • SerpAPI
    • Google Search
    • DuckDuckGo
  • Intelligent webpage scraping and extraction

  • LLM-powered source analysis using Groq

  • Automatic Markdown report generation

  • Research history and report storage

  • Remote MCP deployment support

  • Open-source and extensible architecture

MCP Capabilities

Deep Research MCP provides a complete research workflow through MCP tools and resources:

  • Generate structured research plans
  • Execute iterative web searches
  • Extract and analyze webpage content
  • Gather and compare information from multiple sources
  • Synthesize findings into comprehensive reports
  • Store and retrieve research history
  • Generate Markdown-based research outputs

The server is designed to help AI agents perform deeper, more reliable research by combining search, extraction, analysis, and reporting into a unified workflow.

Architecture

Deep Research MCP follows a modular architecture that includes:

  • Search providers for information discovery
  • Content extraction and scraping components
  • LLM-powered analysis and synthesis
  • Report generation and storage
  • MCP tools and resources for agent interaction

This architecture makes it easy to extend the server with additional search providers, analysis capabilities, and research workflows.

Ver o original no GitHub

Bom para

Relacionados