Skip to main content

Introduction to Airtrain

Airtrain is a powerful platform for building, deploying, and managing AI agents. Whether you're a developer looking to create complex AI systems or a business user needing quick AI solutions, Airtrain provides the tools and infrastructure you need.

What Airtrain Offers​

  • Simplified Agent Creation - Build production-ready AI agents without deep AI expertise
  • Multiple Model Integrations - Connect to OpenAI, Anthropic, Together AI, Fireworks, and more
  • Powerful Abstractions - Execution graphs, permissions systems, and schema validation
  • Deployment Ready - Move from prototype to production with confidence

Core Components​

Agent Framework​

Airtrain's agent framework allows you to create both simple task agents and complex AI employees with persistent memory and decision-making capabilities.

from airtrain.integrations.openai.skills import OpenAIChatSkill, OpenAIInput
from airtrain.integrations.openai.credentials import OpenAICredentials
import os

# Initialize the skill with credentials
credentials = OpenAICredentials(openai_api_key=os.getenv("OPENAI_API_KEY"))
chat_skill = OpenAIChatSkill(credentials=credentials)

# Create input for the model
input_data = OpenAIInput(
    user_input="How do I reset my password?",
    system_prompt="You are a helpful customer support agent.",
    model="gpt-4o",
    temperature=0.7,
    max_tokens=1000,
)

# Get response
result = chat_skill.process(input_data)
print(result.response)

Schema System​

Define precise input and output formats to ensure consistent, reliable agent behavior:

from airtrain.core.schemas import InputSchema, OutputSchema
from pydantic import BaseModel, Field

class TicketInputSchema(InputSchema):
    customer_id: str
    issue_description: str
    priority: int = Field(default=3, ge=1, le=5)

class TicketOutputSchema(OutputSchema):
    ticket_id: str
    assigned_agent: str
    estimated_time: str
    next_steps: list[str]

Getting Started​

Ready to build your first AI agent? Our installation guide will get you set up in minutes.

Or explore our comprehensive Core Principles to understand Airtrain's capabilities in depth.

Use Cases​

  • Customer Support - Create agents that handle customer inquiries and route tickets
  • Content Creation - Deploy agents that generate, edit, and publish content
  • Research Assistance - Build agents that analyze documents and extract insights
  • Process Automation - Develop agents that handle workflows across multiple systems

Airtrain.dev Core Engine​

Airtrain.dev is the core engine that powers the Airtrain platform. As a developer, you get:

  • Fine-grained control over apps, functions, and agents
  • Detailed analysis and debugging capabilities
  • Full visibility into agent behavior and decision-making
  • Customizable agent architectures

Need Help?​

  • Check our documentation
  • Join our GitHub discussions
  • Report issues on GitHub