Skip to main content

Quickstart

This guide will help you install Tyler and create your first AI agent.

Installation​

System requirements​

Tyler requires Python 3.12.8 or later. You can check your Python version by running:

python --version

If you need to install or update Python, visit the official Python website.

Tyler also requires some system libraries for processing PDFs and images. Install them using Homebrew:

brew install libmagic poppler

Installing Tyler​

Install the latest version from PyPI:

pip install tyler-agent

Basic configuration​

Create a .env file in your project directory with your OpenAI API key (minimum required configuration):

# OpenAI API Key (or other LLM provider)
OPENAI_API_KEY=your-api-key-here

For additional configuration options, see the Configuration guide.

Creating your first agent​

Let's create a simple agent that can use web tools and respond to questions. This example demonstrates:

  • Setting up a Tyler agent
  • Creating a conversation thread
  • Sending and receiving messages
  • Using built-in tools
  • Optional Weave monitoring integration
from dotenv import load_dotenv
from tyler.models.agent import Agent
from tyler.models.thread import Thread
from tyler.models.message import Message
import asyncio
import weave
import os
import logging

logger = logging.getLogger(__name__)

# Load environment variables from .env file
load_dotenv()

try:
    if os.getenv("WANDB_API_KEY"):
        weave.init("tyler")
        logger.info("Weave tracing initialized successfully")
except Exception as e:
    logger.warning(f"Failed to initialize weave tracing: {e}. Continuing without weave.")

# Initialize the agent (uses in-memory storage by default)
agent = Agent(
    model_name="gpt-4o",
    purpose="To be a helpful assistant.",
    tools=[
        "web",
        "slack"
    ]
)

async def main():
    # Create a new thread
    thread = Thread()

    # Add a user message
    message = Message(
        role="user",
        content="What tools do you have?"
    )
    thread.add_message(message)

    # Process the thread
    processed_thread, new_messages = await agent.go(thread)

    # Print the assistant's response
    for message in new_messages:
        if message.role == "assistant":
            print(f"Assistant: {message.content}")

if __name__ == "__main__":
    asyncio.run(main())

Understanding the code​

1. Environment setup​

from dotenv import load_dotenv
load_dotenv()

Loads environment variables from your .env file, which should contain your API keys and configuration.

2. Weave monitoring (Optional)​

W&B Weave is a framework for tracking, evaluating, and improving LLM-based applications. While this is optional, you are going to want to use this to understand how your agent is performing.

try:
    if os.getenv("WANDB_API_KEY"):
        weave.init("tyler")
        logger.info("Weave tracing initialized successfully")
except Exception as e:
    logger.warning(f"Failed to initialize weave tracing: {e}. Continuing without weave.")

3. Agent initialization​

agent = Agent(
    model_name="gpt-4o",
    purpose="To be a helpful assistant.",
    tools=[
        "web",
        "slack"
    ]
)

Creates a new Tyler agent with:

  • GPT-4o model
  • A general-purpose role
  • Web and Slack tools enabled
  • Default in-memory storage

4. Thread and message handling​

thread = Thread()
message = Message(
    role="user",
    content="What tools do you have?"
)
thread.add_message(message)

Creates a conversation thread and adds a message to it.

5. Processing and response​

processed_thread, new_messages = await agent.go(thread)

Processes the thread using the agent, which:

  • Analyzes the messages
  • Executes any necessary tools
  • Generates responses

Expected output​

When you run this example, you'll see output similar to:

Assistant: I have access to the following tools:

1. Web Tools:
   - Fetch and process web content
   - Download files from URLs
   - Extract text from web pages

2. Slack Tools:
   - Send messages to channels
   - Upload files
   - Interact with users
   - Manage conversations

I can use these tools to help you with tasks like:
- Looking up information online
- Processing web content
- Communicating through Slack
- Handling file attachments

Is there anything specific you'd like me to help you with using these tools?

Common customizations​

Different model​

agent = Agent(
    model_name="gpt-3.5-turbo",  # Use a faster, cheaper model
    purpose="To be a helpful assistant."
)

Additional tools​

agent = Agent(
    model_name="gpt-4o",
    purpose="To be a helpful assistant.",
    tools=[
        "web",
        "slack",
        "notion",
        "command_line"
    ]
)

Custom system prompt​

thread = Thread(
    system_prompt="You are an AI assistant focused on technical support."
)

Message attributes​

message = Message(
    role="user",
    content="Hello!",
    attributes={
        "source": "website",
        "user_id": "123"
    }
)

Next steps​