Mcp Ai News

AI News Aggregator

Agregador de notícias de IA, papers acadêmicos e repositórios trending de 10 fontes confiáveis.

[](https://mcp-ai-news.onrender.com/health) [](https://python.org) [](LICENSE)

---

Quick Start

Usar API (sem instalação):

curl "https://mcp-ai-news.onrender.com/news?days=7&limit=5"

Instalar MCP para Claude Code:

git clone https://github.com/LauraMattz/mcp_ai_news.git && cd mcp_ai_news
pip install -r requirements.txt
# Recarregue Claude Code (Ctrl+Shift+P → Reload Window)

---

API Pública

Base: https://mcp-ai-news.onrender.com | Docs: Swagger UI

| Endpoint | Descrição | |----------|-----------| | /news?days=7&limit=10 | Todas as notícias | | /search?q=GPT | Busca por palavra-chave | | /papers?days=7 | Papers do ArXiv | | /github?days=7 | Repos trending |

Fontes: OpenAI, Anthropic, Google AI, DeepMind, Hugging Face, MIT Tech Review, TechCrunch, The Verge, ArXiv (cs.AI/cs.LG/cs.CL/cs.CV/stat.ML), GitHub

Features: Cache 15 min • CORS habilitado • Gratuito • ~180+ itens agregados

---

Instalação

MCP Server (Claude Code)

O .mcp.json já está configurado na raiz do projeto. Após git clone e pip install, apenas recarregue Claude Code.

Para Claude Desktop: Adicione ao claude_desktop_config.json (Windows/Mac/Linux paths):

{
  "mcpServers": {
    "ai-news": {
      "command": "python",
      "args": ["/caminho/completo/para/server.py"]
    }
  }
}

REST API (Local)

python api.py
# Acesse: http://localhost:8000/docs

---

Uso

MCP (Claude Code/Desktop):

"Liste as últimas notícias de IA"
"Busque conteúdo sobre transformers"
"Mostre papers do ArXiv sobre computer vision"

API REST:

curl "https://mcp-ai-news.onrender.com/news?days=7&limit=20"
curl "https://mcp-ai-news.onrender.com/search?q=GPT&days=14"
curl "https://mcp-ai-news.onrender.com/papers?days=3&max_results=10"

Documentação completa: Swagger UI • ReDoc

---

Desenvolvimento

Adicionar fonte RSS:

# Em server.py e api.py
FONTES_RSS["nova_fonte"] = {
    "nome": "Nome", "url": "https://...", "ativa": True
}

Testar:

python server.py  # MCP Server
python api.py     # REST API → http://localhost:8000/docs
npx @modelcontextprotocol/inspector python server.py  # MCP Inspector

Estrutura:

├── server.py      # MCP Server (stdio)
├── api.py         # REST API (FastAPI)
├── .mcp.json      # Config Claude Code
└── render.yaml    # Deploy config

---

Licença

MIT License - Ver [LICENSE](LICENSE)

Desenvolvido por Laura Mattos • LinkedIn

Stack: MCP • FastAPI • ArXiv API

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: LauraMattz
• Source: LauraMattz/mcpai_news
• License: MIT
• Homepage: https://mcp-ai-news.onrender.com

Install and usage instructions live in the source repository linked above.