Lightweight Agents (nanobot)

WIP nlp-llms agents automation llm llm-agents mcp nanobot nlp 1 min read

Minimalist, open-source AI agents (nanobot) for tools, chats, and workflows

Overview

While frameworks like LangGraph provide heavy enterprise-grade state machines, lightweight agents focus on extreme modularity, readability, and immediate integration with personal workflows.

nanobot

nanobot (by HKUDS) is an ultra-lightweight, open-source personal AI agent core.

Minimalist Architecture: Written in ~4,000 lines of Python, it keeps the core highly readable while providing real long-running capabilities.
Extensive Integration: Supports 25+ LLM providers, multiple chat channels (Slack, Email, QQ), and acts as a standalone Model Context Protocol (MCP) host.
Practical Use Cases: Excellent for quickly spinning up an agent to handle Web UI chats, background automation (cron jobs), and local tool execution without the overhead of massive frameworks.

TODO: Add deployment snippets and tool integration examples.