LLMilo

Named after Milo, the best dog in the house — LLMilo is an LLM-powered Telegram bot that helps siblings divide and track household chores. Instead of arguing about who was supposed to do the dishes, just ask the bot.

How it works

LLMilo is a Telegram bot backed by an LLM agent with tool-calling capabilities. When you send a message, the agent decides which tools to call to answer your request — whether that's checking what chores are pending, recording that you cleaned the bathroom, or seeing who's been pulling their weight.

The tools are exposed via an MCP (Model Context Protocol) server, so adding new capabilities is as simple as defining a new tool.

Example conversations

• "What chores are pending?" → Shows all pending chores for the household
• "I just mopped the floor" → Records the action
• "Who has done the most chores this week?" → Queries the action history
• "Add a new weekly chore: take out the trash in the kitchen" → Creates the chore

Setup

1. Install dependencies

pip install -r requirements.txt

2. Configure environment

Copy .env.example to .env and fill in the values:

cp .env.example .env

Variable	Description
TELEGRAM_BOT_TOKEN	Bot token from @BotFather
ALLOWED_CHATS	Comma-separated chat IDs the bot will respond to (leave empty to allow all)
LLM_API_BASE	Custom API base URL (for OpenAI-compatible providers)
LLM_API_KEY	API key for the LLM provider
LLM_MODEL	Model identifier (e.g. openai/gpt-4o, anthropic/claude-sonnet-4-20250514)

3. Configure your household

Edit the YAML files in household/:

• household.yaml — Define your household, its rooms/areas (elements), and chores with frequency and optional due times
• people.yaml — Add the people in your household with their names and optional Telegram IDs

4. Run the bot

python -m bot.main

Architecture

LLMilo/
├── agent/          # LLM client and agent loop
│   ├── llm.py      # LiteLLM wrapper for chat completions with tool support
│   └── loop.py     # Agentic loop: call LLM → execute tools → repeat
├── bot/            # Telegram bot interface
│   ├── main.py     # Bot startup, MCP client lifecycle
│   └── handler.py  # Message handler with chat history
├── mcp_tools/      # MCP server & client for tool execution
│   ├── server.py   # FastMCP server exposing household tools
│   └── client.py   # MCP client that connects to the server and bridges tools to LiteLLM
├── household/      # Household data & configuration
│   ├── household.yaml
│   └── people.yaml
├── data/           # SQLite database (auto-created)
├── .env.example
└── requirements.txt

Available tools

Tool	Description
get_household_state	View people, pending chores, and completed chores
write_action	Record that a chore was completed by someone
get_history	Query action history with filters (person, chore, date range)
add_element	Add a room/area to the household
add_chore	Add a chore to an element (daily/weekly/biweekly/monthly/yearly)
add_person	Add a person to the household
remove_element	Remove a room/area and all its chores
remove_chore	Remove a specific chore
remove_person	Remove a person from the household
get_current_time	Get the current date and time

Tech stack

• Python + python-telegram-bot
• LiteLLM for LLM provider abstraction
• MCP for tool server/client
• SQLAlchemy for the persistence layer (SQLite)