Manually Capturing AI Model Events

Capturing Model Data

The library allows you to manually capture the input and output of your AI models using the captureModel function. This is useful when you want to track data outside of HTTP interceptors.

Example

JS

const { captureModel } = require('@handit.ai/node');

// Manually capturing model inputs and outputs
captureModel({
  modelId: 'your-model-id',
  requestBody: { input: 'example input' },
  responseBody: { output: 'example output' }
});

Python

from handit import HanditTracker

tracker = HanditTracker()
tracker.config(api_key="your-api-integration-token")

tracker.capture_model(
    model_id="your-model-id",
    request_body=request_body,
    response_body=response_body,
)

• modelId (required): The unique identifier of the model.

• requestBody (required): The input data sent to the model.

• responseBody (required): The output data generated by the model.

You can locate the model ID of your model below its name.