Nanobot AI

Nanobot is an open-source framework that helps developers build AI agents using the Model Context Protocol (MCP). It focuses on being a flexible “host” that connects AI models to MCP tools and user interfaces, so you can turn models into practical agents without building everything from scratch. It is designed to be readable, extensible, and suitable for real-world applications as well as experimentation.

Key Features

• MCP host for agents: Provides an MCP host so you can plug in MCP servers and turn them into full AI agents.
• MCP-UI support: Works with MCP-UI so agents can show rich interfaces like buttons and views, not just plain text.
• Multi-client ready: Designed to run agents across different clients (such as web or CLI) while keeping shared state and identity.
• Open-source and extensible: Code is public on GitHub under the nanobot-ai organization, making it easy to extend and integrate into your own projects.

Use Cases

• Building custom AI assistants that use MCP tools (for example, code tools, data tools, or company-specific services).
• Adding an MCP-based agent layer to existing products, so they can talk to multiple tools through a common protocol.
• Experimenting with MCP and MCP-UI to prototype new agent behaviors and user experiences.