Overview
Understand AI Dev Kit's agent system architecture and capabilities.
Getting Started
Learn the fundamentals of working with agents:
How Agents Work - Core concepts and architecture
Your First Agent - Quick start guide
Agent vs Request - When to use agents vs direct requests
Core Components
Dive into the agent system's building blocks:
Agent Services - Service architecture and dependencies
Controllers - Conversation, audio, and image controllers
Event Router - Advanced event routing and filtering
Customization
Extend agents with custom implementations:
Custom Chat Services - Integrate custom LLM providers
Architecture
Key Features
Unified API
Interact with any LLM provider using a consistent interface:
OpenAI (Chat Completions, Assistants, Responses, Realtime)
Google Gemini
Anthropic Claude
And more...
Stateful Conversations
Maintain context across multiple interactions:
Automatic history management
Multiple conversation support
Persistent storage options
Multimodal Support
Handle text, voice, images, and more:
Voice input/output
Image generation
Document understanding
Vision capabilities
Tool Integration
Extend agent capabilities with tools:
Built-in tools (search, calculation, etc.)
Custom tool creation
MCP (Model Context Protocol) support
Automatic function calling
Event System
Monitor and respond to agent activities:
Status changes
Streaming updates
Tool execution
Conversation events
Next Steps
Start building with agents:
Create your first agent with Your First Agent
Understand the architecture in How Agents Work
Learn about Agent Services
Explore Event Router for advanced patterns
Last updated