YoMo

YoMo is a self-hosted framework for developers building AI agents that call application tools. It supports distributing inference and tool execution across locations closer to an application's users.

• Agent tools: Developers can write functions in TypeScript or Go that let a model perform actions or retrieve information from an application.
• Tool deployment: The framework handles deployment and management of serverless LLM tools and skills.
• MCP access: You can expose function calls as an MCP server for clients that support the protocol.
• Distributed execution: Inference and tools can run near users in different regions, addressing network delays in interactive applications.

YoMo supports Ollama as a model provider for running models locally. Its tool layer supports different models rather than requiring a single provider. Agent communications use TLS 1.3 encryption.