Home

RubyLLM::Agents Documentation

Welcome to the official documentation for RubyLLM::Agents, a production-ready Rails engine for building, managing, and monitoring LLM-powered AI agents.

Quick Navigation

Getting Started

• Getting Started - Installation and initial setup
• Installation - Detailed installation steps
• Configuration - Configure the initializer
• First Agent - Build your first AI agent
• Generators - Scaffold agents, embedders, and more
• Migration - Upgrade from v0.5.0 to v1.0.0

Core Concepts

• Agent DSL - Declarative agent configuration
• Parameters - Required and optional parameters
• Prompts and Schemas - Structure inputs and outputs
• Conversation History - Multi-turn conversations
• Result Object - Access execution metadata

Features

• Tools - Enable agents to call external functions
• Streaming - Real-time response streaming
• Thinking - Extended reasoning and chain-of-thought
• Attachments - Vision and multimodal support
• Caching - Response caching with TTL
• Execution Tracking - Automatic logging and analytics

Specialized Capabilities

• Embeddings - Text-to-vector embeddings
• Moderation - Content safety filtering
• Audio - Transcription (speech-to-text) and TTS (text-to-speech)
• Image Operations - Generation, analysis, editing, pipelines

Production Features

Reliability

• Reliability Overview - Build resilient agents
• Automatic Retries - Handle transient failures
• Model Fallbacks - Fallback model chains
• Circuit Breakers - Prevent cascading failures

Workflow Orchestration

• Workflows Overview - Compose agents
• Pipeline Workflows - Sequential execution
• Parallel Workflows - Concurrent execution
• Router Workflows - Conditional dispatch

Concurrency

• Async/Fiber - Fiber-based concurrent execution with rate limiting

Governance

• Budget Controls - Spending limits
• Multi-Tenancy - Per-tenant isolation and budgets
• Alerts - Notifications and webhooks
• PII Redaction - Data protection

Development

• Testing Agents - RSpec patterns and mocking
• Error Handling - Error types and recovery

Operations

• Dashboard - Monitoring UI guide
• Database Queries - Execution model queries and analytics
• Production Deployment - Deployment guide
• Background Jobs - Async logging
• Troubleshooting - Common issues

Reference

• Best Practices - Production guidelines
• API Reference - Class documentation
• Examples - Real-world use cases
• FAQ - Common questions
• Contributing - How to contribute

---

About RubyLLM::Agents

RubyLLM::Agents is a Rails engine that provides:

• Clean DSL for defining AI agents with declarative configuration
• Automatic tracking of every execution with costs, tokens, and timing
• Production reliability with retries, fallbacks, and circuit breakers
• Budget controls to prevent runaway costs
• Workflow orchestration for complex multi-agent scenarios
• Real-time dashboard for monitoring and debugging

Current Version

v1.0.0-beta.3 - See CHANGELOG for release history.

Supported LLM Providers

Through RubyLLM, RubyLLM::Agents supports:

• OpenAI - GPT-4, GPT-4o, GPT-4o-mini, GPT-3.5
• Anthropic - Claude 3.5 Sonnet, Claude 3 Opus, Claude 3 Haiku
• Google - Gemini 2.0 Flash, Gemini 1.5 Pro
• And more - Any provider supported by RubyLLM

Need Help?

• GitHub Issues - Report bugs
• GitHub Discussions - Ask questions
• RubyLLM Documentation - Underlying LLM library