AI Text Extractor
Extracting structured information from text.
Node Description
AI Text Extractor - an action type node used to generate a question to AI based on the provided text and receive a response.
Node Configuration
To configure the AI Text Extractor node, you need to fill in the mandatory fields:
- Input Text;
- Entities to Extract.
Input Text
This field is used to enter the text that needs to be analyzed.
Entities to Extract
This field is used to generate a question to the AI and predefine response parameters.
To fill in the Entities to Extract field, you can use the following template:
{
"description": "number of animals mentioned in text", //question to AI
"type": "integer", //format of the attribute with the answer to the question
"var_name": "num_of_animals" //name of the attribute with the answer to the question
}You can extract up to 12 custom request objects from the text. Each entity is described by a JSON with 3-4 key-value pairs:
- Description/description - a string up to 150 characters long for forming the question.
- Type/type - a string for entering the format of the attribute containing the answer. Available formats are integer, string, float, boolean, array.
- Name/var_name - a string up to 50 characters long for entering the name of the attribute containing the answer. The string should start with a letter, followed by letters, numbers, or underscores.
- Valid Values/valid_values - an array to define possible values of the attribute. Mentioning up to 20 values is available. Each value can contain 50 characters.
Example of Node Usage
To obtain an answer to a question based on a given text, you need to create a scenario with the following nodes:
1. Trigger on Webhook to initiate the scenario and provide it with JSON containing the question and response parameters:
[
{
"description": "Number of fruits in text",
"type": "integer",
"var_name": "Answer"
}
]
2. AI Text Extractor to generate an answer to the question and input text, for example, Three apples and four oranges were collected in the garden;
3. Webhook response to receive the answer to the question based on the provided text.
The result of executing the scenario is the answer to the question:
