Build AI Business Solutions with Workflow Automation | NappAI - Template: Conversor de Ticket a JSON

graph TD %%{init: {'theme': 'mc','layout': 'elk'}}%% TextInput-7a156[<div><img src="/_astro/type.Dy26vmDy.svg" style="height: 20px !important;width: 20px !important"/></div>instrucciones] style TextInput-7a156 stroke:#a170ff ChatOutput-ol8bm[<div><img src="/_astro/messages-square.BaSDmT6g.svg" style="height: 20px !important;width: 20px !important"/></div>Chat Output] style ChatOutput-ol8bm stroke:#a170ff OpenAIModel-d3ck8[<div><img src="/_astro/openAI.BhmuxEs3.svg" style="height: 20px !important;width: 20px !important"/></div>OpenAI] style OpenAIModel-d3ck8 stroke:#a170ff TextInput-cv9yq[<div><img src="/_astro/type.Dy26vmDy.svg" style="height: 20px !important;width: 20px !important"/></div>ImagenURL] style TextInput-cv9yq stroke:#a170ff DescribeImage-mc9qi[<div><img src="/_astro/image.segxnpQg.svg" style="height: 20px !important;width: 20px !important"/></div>Describe Image] style DescribeImage-mc9qi stroke:#a170ff ParseData-y6byo[<div><img src="/_astro/braces.Djq0PW4_.svg" style="height: 20px !important;width: 20px !important"/></div>Parse Data] style ParseData-y6byo stroke:#a170ff ParseData-y6byo -.- ChatOutput-ol8bm linkStyle 0 stroke:#a170ff DescribeImage-mc9qi -.- ParseData-y6byo linkStyle 1 stroke:#a170ff TextInput-7a156 -.- DescribeImage-mc9qi linkStyle 2 stroke:#a170ff TextInput-cv9yq -.- DescribeImage-mc9qi linkStyle 3 stroke:#a170ff OpenAIModel-d3ck8 -.- DescribeImage-mc9qi linkStyle 4 stroke:#a170ff

Conversor de Ticket a JSON

🧩 Overview

This workflow automates the extraction of structured data from a fast-food restaurant receipt image. It processes the visual information using an AI model with vision capabilities, guided by specific instructions, and outputs the extracted details in a clean, standardized JSON format. This streamlines data entry and analysis by converting unstructured image data into a machine-readable structure.

⚙️ Main Features

Processes an image of a restaurant receipt to identify and extract key transactional data.
Uses a pre-configured AI model with vision capabilities to analyze the image content.
Applies a detailed set of instructions to guide the AI in extracting specific fields like restaurant name, date, items, and financial totals.
Outputs the extracted data as a structured JSON object, ready for integration with other systems or databases.

🔄 Workflow Steps

Component Name	Role in the Workflow	Key Inputs	Key Outputs
Text Input (Instructions)	Provides the detailed rules and schema for data extraction.	A text block defining the target document type, the list of fields to extract, and the required JSON output structure.	The extraction instructions passed to the AI model.
Text Input (Image URL)	Supplies the source image of the receipt to be processed.	A URL pointing to the digital image of the restaurant ticket.	The image location passed to the vision component.
OpenAI Model	Serves as the core AI engine with vision capabilities for analyzing the image.	Model configuration (e.g., model name, temperature).	A configured language model instance ready for vision tasks.
Describe Image	Analyzes the provided receipt image using the AI model and instructions.	The receipt image URL and the extraction instructions.	A detailed textual description of the image content, structured according to the instructions.
Parse Data	Converts the AI-generated description into a formatted message.	The structured data output from the Describe Image component.	A clean text message containing the extracted data.
Chat Output	Presents the final structured JSON result to the user.	The formatted text message from the Parse Data component.	The final, structured JSON data displayed in the chat interface.

🧠 Notes

The workflow is specifically designed for post-sale receipts from fast-food restaurants; its accuracy may vary with other document types.
The AI model is instructed to output null for any requested field that is not clearly present or identifiable in the receipt image.
The system requires a valid URL for the receipt image to process.
The output is strictly formatted as a JSON object, ensuring compatibility with automated downstream processes.