NudgeLang Documentation

Welcome to the official documentation for NudgeLang, a declarative language for defining autonomous LLM agents.

What is NudgeLang?

NudgeLang is a YAML-based declarative language designed to define autonomous LLM agents in a clear, composable, and maintainable way. It allows you to create agents using well-established agent design patterns from research, with seamless integration of the Model Context Protocol (MCP) for structured context management.

Key Features

• Declarative Syntax: Define agents using clear, readable YAML
• Pattern Support: Implements all common agent patterns
• MCP Integration: Seamless integration with Model Context Protocol
• Tool Management: First-class support for tools and functions
• State Management: Clear state transitions and flow control
• Context Control: Structured approach to context management

Getting Started

1. Installation
2. Quick Start
3. Basic Concepts

Core Components

1. Manifest
2. Models
3. MCP Servers
4. Tools
5. States

Flow Patterns

1. Prompt Chaining
2. Routing
3. Parallelization
4. Orchestrator-Workers
5. Evaluator-Optimizer
6. Autonomous Agent

Best Practices

1. Design Guidelines
2. MCP Integration
3. Security Considerations
4. Performance Optimization

Examples

1. Customer Support Agent
2. Code Review Assistant
3. Content Moderation

API Reference

1. Manifest Schema
2. State Types
3. Tool Definitions
4. MCP Integration