YoMo is an open-source framework for building scalable, ultra-fast AI agents with LLM function calling capabilities. It targets developers building AI-powered applications that require low-latency, secure, and globally distributed inference, aiming to enhance customer experiences through responsive AI interactions.

How It Works

YoMo leverages a geo-distributed architecture and the QUIC protocol for low-latency communication between AI agents and its core server. It emphasizes type-safe function calling for Go and TypeScript, enabling robust agent development with improved error detection and IDE support. The framework simplifies serverless DevOps for LLM tools, allowing developers to focus on agent functionality rather than infrastructure management.

Quick Start & Requirements

• Install CLI: curl -fsSL https://get.yomo.run | sh
• Run server: yomo serve -c my-agent.yaml
• Run agent: yomo run -n get-weather
• Prerequisites: LLM backend (e.g., vLLM, Ollama) with an OpenAI-compatible API endpoint.
• Documentation: yomo.run
• Examples: Serverless LLM Function Calling Examples

Highlighted Details

• Low-latency communication via QUIC protocol.
• TLS v1.3 encryption for enhanced security.
• Type-safe function calling support for TypeScript and Go.
• Geo-distributed architecture for reduced inference latency.

Maintenance & Community

• Active development with a focus on core features.
• Contributing guidelines and a code of conduct are available.

Licensing & Compatibility

• License: Apache License 2.0.
• Compatible with commercial use and closed-source applications.

Limitations & Caveats

The framework is primarily focused on LLM function calling and geo-distributed inference; broader AI model support or advanced orchestration features may require custom integration.