Which integrations are available for this server?

Enables comprehensive Android device automation with 19 tools for device management, screen interaction (tap, swipe, type), screenshots, UI element listing, hardware button simulation, and orientation control via the mobile-mcp server. Provides iOS device control and automation capabilities through the mobile-mcp server, supporting app management, screen interactions, and device control operations. Provides an AI-powered conversational interface through a Telegram bot, enabling natural language interactions for task management, mobile automation, and information retrieval with support for per-user conversation history and chained tool calls.

How do I use Weather Forecast MCP 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., "@Weather Forecast MCP Server what's the forecast for Paris this weekend?" 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.

EasyPomodoro Project Consultant Bot

Telegram bot for consulting on the EasyPomodoro Android project using MCP (Model Context Protocol) servers.

Project Overview

This system provides an AI-powered project consultant that can:

Browse and analyze project code via GitHub Copilot MCP (HTTP transport)
Search project documentation using RAG (Retrieval Augmented Generation)
Explore project structure with tree navigation

Architecture

┌─────────────┐ │ Telegram │ │ User │ └──────┬──────┘ │ ↓ ┌─────────────────────────────────────────┐ │ Telegram Bot (client/bot.py) │ │ - Handles user messages │ │ - Manages conversation history │ │ - Shows "Думаю..." indicator │ │ - Filters tools to essential set │ └──────┬──────────────────────────────────┘ │ ↓ ┌─────────────────────────────────────────┐ │ OpenRouter API │ │ Model: deepseek/deepseek-v3.2 │ │ - Processes natural language │ │ - Decides when to use tools │ └──────┬──────────────────────────────────┘ │ (when tool needed) ↓ ┌─────────────────────────────────────────┐ │ MCP Manager (mcp_manager.py) │ │ - Manages 2 MCP server connections │ │ - Routes tool calls to correct server │ │ - HTTP transport for GitHub Copilot │ │ - stdio transport for RAG MCP │ └──────┬──────────────────────────────────┘ │ ├───────────────────────────────────┐ ↓ ↓ ┌──────────────────────┐ ┌──────────────────────┐ │ GitHub Copilot MCP │ │ RAG Specs MCP │ │ (HTTP Transport) │ │ (Python/stdio) │ │ │ │ │ │ URL: │ │ Tools: │ │ api.githubcopilot. │ │ - rag_query │ │ com/mcp/ │ │ - list_specs │ │ │ │ - get_spec_content │ │ Essential Tools: │ │ - rebuild_index │ │ - get_file_contents │ │ - get_project_ │ │ - list_commits │ │ structure (tree) │ │ - get_commit │ │ │ │ - list_issues │ │ Uses: │ │ - issue_read │ │ - GitHub API │ │ - list_pull_requests │ │ - FAISS + Ollama │ │ - pull_request_read │ │ │ └──────────────────────┘ └──────────────────────┘

System Components

1. MCP Servers

1.1 GitHub Copilot MCP (HTTP)

Purpose: Provide access to GitHub repository via GitHub Copilot's MCP endpoint

URL: https://api.githubcopilot.com/mcp/

Transport: HTTP (Streamable HTTP transport, MCP spec 2025-03-26)

Essential Tools (filtered for token efficiency):

get_file_contents - Read file contents from repository
list_commits / get_commit - View commit history
list_issues / issue_read - Work with issues
list_pull_requests / pull_request_read - Work with PRs

Note: The server provides 40+ tools, but only essential ones are sent to the model to reduce token usage.

Authentication: GitHub Personal Access Token (PAT)

1.2 RAG Specs MCP (Python)

Purpose: Search project documentation using RAG and explore project structure

Location: mcp_rag/

Files:

server.py - MCP server with RAG tools
github_fetcher.py - GitHub API client for /specs folder and project structure
rag_engine.py - FAISS + Ollama embeddings

Tools:

rag_query - Search documentation with semantic similarity
list_specs - List available specification files
get_spec_content - Get full content of a spec file
rebuild_index - Rebuild the RAG index
get_project_structure - Get directory tree (use FIRST to find file paths)

Target Repository: LebedAlIv2601/EasyPomodoro Specs Path: /specs

2. Telegram Bot Client (client/)

Files:

main.py - Application entry point
bot.py - Telegram bot handlers with tool call loop
mcp_manager.py - MCP server management (HTTP + stdio)
mcp_http_transport.py - HTTP transport for GitHub Copilot MCP
openrouter_client.py - OpenRouter API integration
conversation.py - Per-user conversation history
logger.py - Logging configuration
config.py - Configuration, environment variables, and ESSENTIAL_TOOLS filter

Installation

Prerequisites

Python 3.14+
Ollama with nomic-embed-text model (for RAG)
Telegram bot token
OpenRouter API key
GitHub Personal Access Token

Setup

Clone repository:

cd /path/to/McpSystem

Create virtual environment:

python3.14 -m venv venv source venv/bin/activate

Install dependencies:

pip install -r requirements.txt pip install -r mcp_rag/requirements.txt

Install Ollama model (for RAG):

ollama pull nomic-embed-text

Configure environment:

cd client cp .env.example .env # Edit .env with your credentials: # TELEGRAM_BOT_TOKEN=your_token # OPENROUTER_API_KEY=your_key # GITHUB_TOKEN=your_github_pat

GitHub PAT Scopes

Create a Classic PAT with these scopes:

repo - Full repository access
read:org - Read organization data (optional)
read:user - Read user data

Running

cd client ../venv/bin/python main.py

The bot will:

Connect to GitHub Copilot MCP (HTTP)
Start RAG Specs MCP server (Python/stdio)
Fetch and filter tools to essential set (~12 tools)
Start Telegram bot polling

Usage

Commands

/start - Show welcome message

Example Queries

Documentation questions:

"What is the project architecture?"
"How does the timer feature work?"
"List all specification files"

Code questions:

"Show me the main activity code"
"What files are in the app module?"
"Get project structure"

GitHub questions:

"Show recent commits"
"List open issues"
"What pull requests are pending?"

Recommended Workflow

For code exploration, the model follows this workflow:

get_project_structure - Find file paths first
get_file_contents - Read specific files using exact paths

Configuration

Environment Variables

Variable	Description
`TELEGRAM_BOT_TOKEN`	Telegram bot token
`OPENROUTER_API_KEY`	OpenRouter API key
`GITHUB_TOKEN`	GitHub Personal Access Token

config.py Settings

Setting	Default	Description
`OPENROUTER_MODEL`	`deepseek/deepseek-v3.2`	AI model
`GITHUB_OWNER`	`LebedAlIv2601`	Repository owner
`GITHUB_REPO`	`EasyPomodoro`	Repository name
`SPECS_PATH`	`specs`	Documentation folder
`MAX_CONVERSATION_HISTORY`	50	Max messages per user
`TOOL_CALL_TIMEOUT`	120.0	MCP tool timeout (seconds)
`ESSENTIAL_TOOLS`	list	Tools to send to model (token optimization)

Technology Stack

Python 3.14 - Main language
python-telegram-bot - Telegram integration
MCP SDK - Model Context Protocol (HTTP + stdio transports)
httpx - Async HTTP client for GitHub Copilot MCP
FAISS - Vector similarity search
Ollama - Local embeddings (nomic-embed-text)
OpenRouter - AI model access

Project Statistics

MCP Servers: 2
- GitHub Copilot MCP (HTTP, ~40 tools available, ~8 essential)
- RAG Specs MCP (Python/stdio, 5 tools)
Essential Tools: ~12 (filtered for token efficiency)

Troubleshooting

GitHub Copilot MCP connection errors

Verify PAT has correct scopes (repo, read:org)
Check token is not expired
Ensure token is in .env file
Check network connectivity to api.githubcopilot.com

RAG not working

Verify Ollama is running: curl http://localhost:11434/api/tags
Check nomic-embed-text model: ollama list

High token usage

Ensure ESSENTIAL_TOOLS filter is applied in config.py
Check logs for "Filtered tools: X/Y" message

License

This project demonstrates MCP integration for AI-powered project consultation.

Weather Forecast MCP Server