Cosa Sai

🚀 MCP Server for Technical Document Access

This project implements an MCP server that utilizes the Gemini API to provide access to various technical documents, specifically tailored for the Roo/Cline environment. By leveraging Gemini's large context window, users can explore and understand different technologies more efficiently.

🚀 Quick Start

The MCP server offers a convenient way to access curated technical documents. With the power of the Gemini API, it overcomes the limitations of traditional context windows and supports a wide range of technologies.

✨ Features

Advantages

1. Access to a curated knowledge base: Users can directly obtain curated technical documents without the need to collect information on their own.
2. Overcome context window limitations: The powerful capabilities of the Gemini API enable us to handle larger document contents, improving query efficiency.
3. Support for multiple technologies: The server can be extended to support more different technical domains.

Functionality

• Supports multiple query tools:
  • can_x_be_done: Checks if a specific task can be implemented in a specified technology.
  • hints_for_problem: Gets hints and suggestions for solving problems.
  • is_this_good_practice: Evaluates whether a code snippet follows good programming practices.
  • how_to_do_x: Provides methods and alternatives for completing a specific task.

📦 Installation

Installation via Smithery

gem install gemini-docs-mcp-server

Configuration Example

{
  "server": {
    "host": "localhost",
    "port": 3000,
    "apiKey": "your_api_key_here"
  },
  "tools": [
    {
      "name": "can_x_be_done",
      "enabled": true
    }
  ]
}

💻 Usage Examples

Knowledge Base Retrieval and Cleaning

Web Scraping Script

wget --mirror --convert-links --adjust-extension --page-requisites --no-parent --directory-prefix=./local_copy --no-verbose --show-progress $1

Document Conversion Tool

#!/bin/bash

directory="${1:-.}"  # Default to current directory if no argument is provided
output_file="${2:-concatenated.md}"  # Default output file name

echo "Merging files in '$directory' into '$output_file'..."
# Empty the output file if it exists
truncate -s 0 "$output_file"

# Find all files (excluding directories) and process them
find "$directory" -type f -name '*.html' | while IFS= read -r file; do
    echo "=== ${file#./} ===" >> "$output_file"
    cat "$file" \
    | grep -v 'base64' \
    | html2markdown >> "$output_file"
    echo -e "\n" >> "$output_file"
done

echo "Done! Output saved to '$output_file'"

Tool Details

• can_x_be_done: Checks if a task can be implemented in a specified technology.

  • Input: docs, prompt, x, technology
  • Output: success, data

• hints_for_problem: Gets hints for solving a problem.

  • Input: docs, prompt, problem, context, environment
  • Output: success, data

• is_this_good_practice: Evaluates the quality of a code snippet.

  • Input: docs, prompt, snippet, context
  • Output: success, data

• how_to_do_x: Gets methods for completing a task.

  • Input: docs, prompt, x, technology
  • Output: success, data

🤝 Contribution Guidelines

1. Create a feature branch: Create a new branch for new features or bug fixes.
2. Submit code: Ensure that the commit message is clearly described for easy tracking of changes.
3. Submit a pull request: Merge your changes into the main branch.

📄 License

This project is licensed under the MIT License.

⚠️ Disclaimer

This version is still in its early stages and may have errors and limitations. If you encounter any problems, please report them at any time, and suggestions for improvement are welcome.