Weave

Integrating Kern with Weave by WandB

Weave by Weights & Biases (WandB) provides a powerful platform for logging and visualizing model calls. By integrating Kern with Weave, you can track and analyze your agent's performance and behavior.

Prerequisites

1. Install Weave

   Ensure you have the Weave package installed:

   1uv pip install weave

2. Authentication Go to WandB and copy your API key

   1export WANDB_API_KEY=<your-api-key>

Logging Model Calls with Weave

This example demonstrates how to use Weave to log model calls.

1import weave
2from kern.agent import Agent
3from kern.models.openai import OpenAIResponses
4
5# Initialize Weave with your project name
6weave.init("kern")
7
8# Create and configure the agent
9agent = Agent(model=OpenAIResponses(id="gpt-5.2"), markdown=True, debug_mode=True)
10
11# Define a function to run the agent, decorated with weave.op()
12@weave.op()
13def run(content: str):
14    return agent.run(content)
15
16# Use the function to log a model call
17run("Share a 2 sentence horror story")

Notes

• Environment Variables: Ensure your environment variable is correctly set for the WandB API key.
• Initialization: Call weave.init("project-name") to initialize Weave with your project name.
• Decorators: Use @weave.op() to decorate functions you want to log with Weave.

By following these steps, you can effectively integrate Kern with Weave, enabling comprehensive logging and visualization of your AI agents' model calls.