Extract data to a CSV using AI.

Node Inputs

  1. Text for extraction: Text to extract table from
    • Type: Text
  2. Additional Context: Provide any additional context that may help the AI model extract the data.
    • Type: Text
  3. Select CSV: Your file should contain headers in the first row.
    • Type: File

For providing the CSV file, you have two options:

  1. If toggle is on Upload: Upload a file by clicking the upload button
  2. If toggle is on Variable: Reference files from other nodes.

If Add columns manually is selected:

  1. Column Name: The name of the column.
    • Type: Text
  2. Column Value: The value of the column.
    • Type: Text

Node Parameters

On the face of the node:

  1. Extract multiple rows: Choose the mode of extraction. If checked, it will extract multiple rows of data. If unchecked, it will extract a single row.
    • Type: Toggle
  2. Add columns manually: Add data points for some columns manually instead of having them extracted by the AI model.
    • Type: Toggle

In the gear:

  1. Provider: The model provider.
    • Type: Dropdown
  2. Model: The specific model for extracting the table
    • Type: Dropdown

Node Outputs

  1. Table: The table extracted from the text.
    • Type: File
    • Example usage: {{extract_to_table_0.table}}

Example

The below example shows a pipeline that extracts employee information from text and saves it as a CSV file.

  1. Text Node: Contains the text with employee information.
    • Text: Our company has 3 employees. John Smith is our CEO and earns $120,000 per year. Sarah Johnson is our CTO with a salary of $110,000. Mike Brown is our Marketing Manager making $90,000 annually.
  2. Input Node: Contains the CSV file.
    • Type: File
  3. Extract to CSV Node: Extracts the data and saves it as CSV.
    • Text for extraction: {{text_0.text}}
    • Additional Context: Please extract employee information including name, role, and salary.
    • Select CSV: {{input_0.file}}
    • Extract multiple rows: True
    • Add columns manually: False
  4. Output: The extracted CSV file
    • Output: {{extract_to_csv_0.table}}