Which integrations are available for this server?

Converts scraped web pages into a clean Markdown format, optimizing the content for easy ingestion and processing by AI agents.

How do I use FastMCP Documentation Search Server?

1. Click on "Install Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@FastMCP Documentation Search Server search for how to create a new FastMCP server" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

FastMCP Search Server 🚀

Português | English

Português

Servidor baseado no protocolo MCP (Model Context Protocol) projetado para fornecer uma infraestrutura de Arquitetura de Acesso + Contexto. Este sistema permite que Agentes de IA estendam suas capacidades através de ferramentas locais e recuperação de dados especializados sem a necessidade de processamento de LLM no lado do servidor.

🏗️ Arquitetura e Funcionamento

O sistema opera como uma camada intermediária de inteligência local, automatizando a busca e o processamento de dados para injetar apenas o necessário na janela de contexto do cliente.

graph TD User((Usuário)) --> Client[MCP Client / Interface] subgraph "Camada de Comunicação" Client <==> Protocol(MCP Protocol) end subgraph "FastMCP Server (Infraestrutura Local)" Protocol <==> Tools{Motor de Ferramentas} Tools --> Index[minsearch / TF-IDF] Tools --> Scraping[Jina Reader] Tools --> Logic[Lógica Local] end Index --- Docs[(Documentação Local)] Scraping --- Web((Web))

🛠️ Ferramentas, Inputs e Outputs

Ferramenta	Descrição	Input	Output
`search_docs`	Busca semântica inteligente usando TF-IDF.	`query` (string)	Lista dos 5 documentos mais relevantes com preview.
`scrape_page`	Web scraping otimizado para IA.	`url` (string)	Conteúdo da página em Markdown limpo.
`hash_text`	Geração de hash para integridade.	`text` (string)	String SHA-256 hexadecimal.
`add`	Operação aritmética precisa.	`a` (int), `b` (int)	Soma literal dos números.

💻 Stack Tecnológica

FastMCP: Framework principal para orquestração do protocolo.
minsearch: Motor de busca minimalista para indexação in-memory.
Scikit-learn & Pandas: Vetorização e manipulação de dados estruturados.
Jina Reader API: Conversão de HTML para Markdown legível por IA.

🚀 Instalação

# Clone o repositório e instale as dependências uv sync # Execute o servidor uv run python main.py

English

A server based on the Model Context Protocol (MCP) designed to provide an Architecture of Access + Context. This system allows AI Agents to extend their capabilities through local tools and specialized data retrieval without the need for LLM processing on the server side.

🏗️ Architecture and Workflow

The system operates as an intermediate layer of local intelligence, automating data search and processing to inject only what is necessary into the client's context window.

graph TD User((User)) --> Client[MCP Client / Interface] subgraph "Communication Layer" Client <==> Protocol(MCP Protocol) end subgraph "FastMCP Server (Local Infrastructure)" Protocol <==> Tools{Tools Engine} Tools --> Index[minsearch / TF-IDF] Tools --> Scraping[Jina Reader] Tools --> Logic[Local Logic] end Index --- Docs[(Local Docs)] Scraping --- Web((Web))

🛠️ Tools, Inputs, and Outputs

Tool	Description	Input	Output
`search_docs`	Intelligent semantic search using TF-IDF.	`query` (string)	List of the 5 most relevant docs with content preview.
`scrape_page`	AI-optimized web scraping.	`url` (string)	Page content in clean Markdown.
`hash_text`	Hash generation for data integrity.	`text` (string)	SHA-256 hexadecimal string.
`add`	Precise arithmetic operation.	`a` (int), `b` (int)	Literal sum of the numbers.

💻 Technical Stack

FastMCP: Core framework for protocol orchestration.
minsearch: Minimalist search engine for in-memory indexing.
Scikit-learn & Pandas: Vectorization and structured data handling.
Jina Reader API: HTML to AI-readable Markdown conversion.

🚀 Getting Started

# Clone the repository and install dependencies uv sync # Run the server uv run python main.py

📝 Conclusão / Conclusion

Este projeto demonstra a viabilidade de construir camadas de suporte para agentes de IA que priorizam a eficiência e a soberania dos dados. Ao utilizar o protocolo MCP, removemos a fricção entre bases de dados locais e modelos globais, garantindo que o contexto injetado seja preciso, relevante e processado de forma otimizada.

This project demonstrates the feasibility of building support layers for AI agents that prioritize efficiency and data sovereignty. By using the MCP protocol, we remove the friction between local databases and global models, ensuring that the injected context is accurate, relevant, and optimally processed.

Developed as part of the AI Dev Bootcamp.