AI/TLDR

nanobot

Self-hosted personal agent runtime with a WebUI, memory, MCP and chat apps

Agent Frameworks & BuildersOpen source
Language
Python
License
MIT
$uv tool install nanobot-ai

Overview

nanobot is a lightweight, self-hosted personal AI agent framework written in Python. It bundles the pieces a long-running assistant needs — tools, long-term memory, MCP integrations, model routing, multi-agent delegation, scheduled automation and an OpenAI-compatible API — into a small, readable core that you run on your own machine or server.

You can talk to it three ways: a bundled browser WebUI that binds to localhost on first run, an interactive terminal chat, or chat apps such as Telegram, Discord, Slack, WeChat, Feishu, Teams, Mattermost and email. The gateway can be left running in the background so channels and scheduled automations keep working after you close the terminal.

The agent's tool set covers files, shell, web search, web fetch, MCP servers, cron, image generation and sub-agents, and session history plus long-term memory are handled by its Dream memory layer. Providers are configured in the WebUI: any OpenAI-compatible API, local models through Ollama or vLLM, and fallback models when one provider fails. It is MIT-licensed, needs Python 3.11 or newer, and is published to PyPI as nanobot-ai.

What it does

  • Runs as a browser WebUI, an interactive terminal chat, or inside chat apps (Telegram, Discord, Slack, WeChat, Feishu, Teams, email, Mattermost)
  • Built-in tools: files, shell, web search, web fetch, MCP, cron, image generation and sub-agents
  • Long-term memory and session history via its Dream memory layer, for work that spans sessions
  • Model freedom: OpenAI-compatible endpoints, local LLMs, and configurable fallback models
  • Python SDK plus an OpenAI-compatible API, so other apps can call the agent
  • Runs as a long-lived background gateway with status, logs, restart and stop commands

Getting started

Install the nanobot-ai package, start the WebUI, and pick a provider and model in Settings before doing real work.

Install nanobot

Install the published package with uv (Python 3.11 or newer). The README also offers a one-command installer script for macOS, Linux and Windows PowerShell.

bashbash
uv tool install nanobot-ai

Open the WebUI and configure a model

The recommended first run. It creates the config and workspace, starts the gateway, and opens http://127.0.0.1:8765. Then go to Settings → Models, choose a provider, credential and model, and send "Hello!" in a new topic to check the connection.

bashbash
nanobot webui

Or stay in the terminal

nanobot agent opens an interactive terminal chat with the same model, workspace and tools. The -m form sends one request and exits — handy for scripts or a quick provider check.

bashbash
nanobot agent
nanobot agent -m "Hello!"

Keep it running in the background

Once first-time model setup is done, run the gateway detached and manage it with the gateway subcommands.

bashbash
nanobot webui --background
nanobot gateway status
nanobot gateway logs
nanobot gateway stop

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

When to use it

  • You want one personal assistant you own end to end, running on your own hardware with your own keys
  • You want the assistant reachable from a chat app you already use rather than only a web page
  • You need scheduled or long-horizon automations that keep running between conversations
  • You want an agent that carries memory across sessions and can call MCP servers and shell tools

How nanobot compares

nanobot alongside other open-source agent frameworks & builders tools AI/TLDR tracks, ranked by GitHub stars.

ToolStarsWhat it does
DeepSeek Harness★ 209kDeepSeek AI's open-source agent harness (dsh), built on Cordis, where models, tools, skills, sessions, sandboxes, storage and the UI are all plugins composed through profiles.
AutoGPT★ 187kOne of the earliest autonomous agent projects, now a platform for building and running agents from reusable blocks and workflows.
DeerFlow★ 81.3kByteDance's open-source super agent harness built on LangGraph: skills, sub-agents, sandboxes, a filesystem and long-term memory for long-horizon research, coding and content tasks.
nanobot★ 47.7kSelf-hosted personal agent runtime with a WebUI, memory, MCP and chat apps
Agno★ 42kA fast Python framework (formerly Phidata) for building agents with memory, tools, and multimodal inputs, plus a runtime for deploying them in production.
LangGraph★ 40.9kA library from the LangChain team for building stateful, graph-based agent workflows with explicit control over steps, memory, and human-in-the-loop checkpoints.
AgentGPT★ 36.3kAgentGPT lets you name a custom AI, give it a goal, and watch it plan tasks, run them, and learn from the results, all from a web browser.
STORM★ 31.2kStanford OVAL's LLM-powered knowledge curation system that researches a topic and generates a Wikipedia-style article with citations.