Skip to content
Nappai Documentation

Categorizer

This component helps you quickly organize your data into categories. It uses advanced AI to understand the content of your data and group similar items together. This makes it easier to find and work with your information.

Relationship with AI Language Models

This component uses an AI language model (like those from OpenAI) to understand the meaning of your data and determine the best categories. You’ll need to select the language model you want to use.

Inputs

  • Items: This is the data you want to categorize. It can be text, numbers, or other data types. This field is required.
  • Max Chunks: This sets a limit on how much data the AI processes at once. A higher number might be more accurate but will take longer. The default is 5. You can adjust this if needed.
  • Language Model: You need to select the specific AI language model you want to use for categorization. This field is required.

Outputs

  • Categories: This is a list of the categories the AI has identified in your data. You can use this list in other parts of Nappai to filter, sort, or analyze your data further.
  • Tool: This provides a reusable tool that performs the categorization. This is more advanced and might be useful for developers or users who want to integrate this functionality into custom workflows.

Usage Example

Imagine you have a list of customer reviews. You can feed these reviews into the “Categorizer” component. The component will then analyze the reviews and output categories like “Positive Feedback,” “Negative Feedback,” and “Feature Requests.” You can then use these categories to filter the reviews and focus on specific areas of interest.

Templates

[List of templates where the component can be seen and its configuration - This section requires information not provided in the original prompt. Please provide template information to complete this section.]

  • Summarizer: Use this component to shorten your data before categorizing it, especially if you have very long texts.
  • Entities extraction: This component can extract key information from your data before sending it to the Categorizer for more precise categorization.
  • Data Conditional Router: Route the categorized data to different parts of your workflow based on the categories assigned.
  • OpenAI Function Agent: The underlying AI model used by the Categorizer is often an OpenAI model.

Tips and Best Practices

  • Start with a smaller amount of data to test the categorization.
  • Experiment with different language models to see which one works best for your data.
  • Adjust the “Max Chunks” setting to balance accuracy and processing time.
  • Consider pre-processing your data (e.g., cleaning or formatting) before using the Categorizer for better results.

Security Considerations

  • Ensure the language model you select is appropriate for the sensitivity of your data.
  • Review the privacy policies of the language model provider. Nappai will not store your data beyond what is necessary for processing. However, the underlying AI model provider may have its own data retention policies. Please consult their documentation.