What is nanobot?
nanobot is an open-source, self-hosted personal AI agent framework written in Python. nanobot provides a browser, terminal, or chat app interface for working with files, remembering preferences, and scheduling tasks. The project is available from the HKUDS GitHub repository under the MIT license and connects to an external model provider or local model server.
nanobot combines personal assistant workflows with developer-facing integration options. Saved conversations, long-term memory, and task results support work that continues across sessions. The repository also documents deployment as a long-running local or server-side agent gateway, allowing users to inspect, customize, and extend their own installation.
What nanobot can do
nanobot brings document handling, automation, memory, and external tools into the same agent framework. Built-in file tools support report summarization and other document tasks, while chat channels provide access outside the browser. Additional capabilities include shell operations, web tools, image generation, and delegation to multiple agents.
- nanobot reads PDF, Word, Excel, and PowerPoint files with built-in file tools.
- nanobot keeps session history and long-term memory through Dream.
- nanobot runs scheduled automations, cron tasks, and long-horizon goals.
- nanobot uses shell tools, web search, and web fetching.
- nanobot supports MCP integrations, image generation, and multi-agent delegation.
- nanobot exposes a Python SDK and an OpenAI-compatible API.
How nanobot works
nanobot requires Python 3.11 or newer and access to a supported model provider or local model server. The recommended stable installer selects an isolated Python environment and opens setup. Users who manage their own environments can install the nanobot-ai package with uv or pip, while developers can install an editable source checkout using Git and Bun.
nanobot browser setup starts with the nanobot webui command when setup has not already opened. Users save a provider and model preset in Settings → Models, then start a topic to check the connection. File work uses a selected project folder and permissions; attaching a file to a conversation does not change that selected folder.
Who nanobot is for
nanobot serves people who want a personal assistant they can self-host, along with developers who need programmatic agent access. Browser and terminal interfaces support direct use, while Telegram, Discord, Slack, WeChat, Email, Mattermost, and Linear connections support chat-based workflows. The documentation includes a starting path for users without terminal or configuration experience.
- nanobot can summarize reports using supported document formats.
- nanobot can read notes and write a summary in a selected workspace.
- nanobot can retain preferences and task results for later work.
- nanobot can run scheduled tasks and ongoing goals.
- nanobot can connect applications through Python or HTTP integrations.
nanobot pricing and access
nanobot is distributed as an MIT-licensed open-source project through a stable installer, the nanobot-ai Python package, or a source checkout. The repository does not supply a model account. Model fees or local hardware requirements depend on the provider or local model server selected, so the software license does not remove those separate operating requirements.
nanobot strengths and limitations
nanobot combines self-hosting, multiple interfaces, persistent workflows, and model choice in a customizable framework. The stable release includes the browser interface, and supported platform wheels also bundle the native terminal interface. The repository distinguishes stable releases for everyday use from current source for development and newer features, making the installation track an important setup choice.
- nanobot requires a separately configured model provider or local model server.
- nanobot source installation requires Git and Bun.
- nanobot native terminal interface support depends on the operating system and runtime.
- nanobot users on other platforms must use the classic terminal mode or WebUI.
nanobot FAQ
What is nanobot?
nanobot is an open-source, self-hosted personal AI agent framework written in Python. nanobot works through a browser, terminal, or chat app to handle files, remember preferences, and schedule tasks, using a supported model provider or local model server.
Is nanobot free?
nanobot is available as MIT-licensed open-source software. nanobot does not supply a model account, and model fees or local hardware requirements depend on the service chosen. Access is available through a stable installer, Python package, or source checkout.
What can you use nanobot for?
nanobot can summarize documents, work with workspace files, remember preferences, and run scheduled automations. nanobot also supports shell tools, web search and fetching, image generation, MCP integrations, and multi-agent delegation for broader agent workflows.
Does nanobot have an API or integrations?
nanobot exposes a Python SDK and an OpenAI-compatible API for integrations. nanobot also supports MCP and connects to communication channels including Telegram, Discord, Slack, WeChat, Email, Mattermost, and Linear.
What do you need to run nanobot?
nanobot requires Python 3.11 or newer and access to a supported model provider or local model server. nanobot offers a stable installer and package installation options; installing the current source additionally requires Git and Bun.
Can nanobot run locally?
nanobot supports self-hosted operation and deployment as a long-running local or server-side agent gateway. nanobot can connect to a local model server, and users can access the agent through the bundled browser interface or terminal.
From the community
No reviews yet
Be the first to review this tool. What you write helps the next person choose.




