Run a personal AI agent that reads files and schedules tasks from your browser or chat apps
Installing an AI assistant on your own computer or server lets you chat via a browser or apps like Telegram to summarize PDFs, organize files, and run tasks at specific times. nanobot is a lightweight Python framework that runs personalized AI by connecting to model providers or local models without relying on external clouds. It stores conversation history and settings to maintain context from previous work.
Changes in v0.4.0
The recently released v0.4.0 version adds the ability to permanently store conversation history in SQLite. You can retrieve existing conversation records and maintain source backups, making it easy to restore previous work content even if a session is interrupted. The update also improves the ability to view Mermaid diagrams in the terminal and connect to remote instances. This update focuses on helping you continue daily tasks without interruption.
Channel Integration and File Handling
nanobot can connect to a web UI, terminal, and various chat platforms such as Telegram, Discord, Slack, WeChat, and email.
- Chat App Integration You can connect bots to various channels including Telegram, Discord, Slack, WeChat, email, Mattermost, and Linear.
- File Tools It performs tasks such as reading, summarizing, or converting PDF, .docx, .xlsx, and .pptx files.
- Automation and Scheduling You can use cron expressions to run tasks at specific times or set long-term goals for the agent to continue working on.
- Model Flexibility It supports connecting to local model servers such as OpenAI-compatible APIs, Ollama, and vLLM, as well as image generation and web search tools.
Installation and Runtime Environment
Python 3.11 or higher is required, and you must have access to a supported model provider account or a local model server. Running the installation script automatically creates an isolated Python environment and starts model configuration via the web UI. You can also install the nanobot-ai package using uv or pip. If you want the latest features, you must clone the repository via Git and install it in editable mode, which requires Bun. After installation, complete the setup in your browser using the nanobot webui command.
Considering Self-Hosting
nanobot is suitable for those who want a personal AI agent that they can fully control on their own infrastructure. It is useful for developers or power users who want to manage conversation history and memory directly or integrate specific chat apps with file systems. However, model API costs or local hardware requirements vary depending on the service you choose, and nanobot itself does not provide model accounts. You must run a gateway in the background, and automatic startup after a system reboot must be configured separately through operating system service settings. You must always check file access permissions and security settings before handling important data.