nanobot

TLDR

Start an interactive session

$ nanobot

Send a one-off command

$ nanobot "[task]"

Run with a specific local model

$ nanobot --model [model_name]

Start in server mode for Telegram integration

$ nanobot serve --telegram

Schedule a recurring task

$ nanobot schedule "[task]" --cron "[cron_expression]"

List active scheduled tasks

$ nanobot tasks

SYNOPSIS

nanobot [options] [command]
nanobot serve [options]
nanobot schedule [options]

nanobot is an ultra-lightweight open-source AI assistant built in approximately 4,000 lines of code. It supports interaction through messaging platforms including Telegram and WhatsApp, enabling voice and text-based communication from any device.
The assistant runs local language models for privacy-focused operation and can execute scheduled tasks autonomously using cron expressions. It supports multi-step task automation, reminders, and conversational interaction without requiring cloud AI services.
nanobot is designed to be simple to deploy and extend, with minimal dependencies and a small footprint suitable for running on low-resource hardware including Raspberry Pi devices.

PARAMETERS

COMMAND

Natural language task or instruction for the AI assistant.

serve

Start nanobot in server mode for messaging platform integration.

--telegram

Enable Telegram bot integration.

--whatsapp

Enable WhatsApp integration.

schedule

Create a scheduled or recurring task.

--cron EXPRESSION

Cron expression for task scheduling.

tasks

List all active scheduled tasks.

--model MODEL

Specify the local LLM to use.

--local

Force fully local operation without external APIs.

--port PORT

Port number for server mode (default: 3000).

--verbose

Enable verbose logging output.

--help

Display help information.

--version

Display version information.

CAVEATS

Messaging platform integrations require API tokens from their respective services. Running local models requires sufficient RAM and compute resources. Scheduled tasks persist only while the server process is running unless configured with a process manager.

HISTORY

nanobot was created by the HKUDS research group (Data Intelligence Lab at the University of Hong Kong). It was designed as a minimalist alternative to larger AI assistant frameworks, emphasizing simplicity and local-first operation. The project gained attention for achieving broad AI assistant functionality in a remarkably small codebase.