Azure Document Intelligence Web/Desktop App
Budget: €250 – €750 EUR
Develop a Web and Desktop Application using Azure Document Intelligence (Azure Form Recognizer) to recognize and process invoice formats.
1. Integration of Azure Document Intelligence with the Application
Configure Azure Document Intelligence:
Create an Azure Form Recognizer resource in the Azure portal.
Use its “Custom Model” functionality to train the AI with specific invoice examples (at least 5 for each format).
Export the model and get the API credentials (Endpoint and Key).
2. Application Development:
Create a module that allows uploading invoices (PDF, images) from the backend, portal, directory, email, etc .
Use Azure credentials to send documents to the API and receive processed responses.
3. Process the received information:
Map the extracted fields (invoice number, date, total, tax, vendor, etc.) to the prefixed field templates in the application.
Add logic to validate and save this data in the corresponding records.
2. Application Functionalities
Upload invoices: Allow users to upload documents manually or process them automatically from emails or shared folders.
Integration with Azure API: Send the uploaded documents to the Azure API and receive the extracted data.
Visualization and editing: Display the extracted data in an editable form before confirming the registration in the application.
Additional training: A simple system within the application to add new invoice examples that the AI does not recognize well and retrain the model in Azure.
3. Application Flow
1. User loads or selects an invoice.
2. Application sends the invoice to Azure Document Intelligence.
3. Azure returns the extracted data.
4. The data is automatically mapped into an invoice form in the Application.
5. The user reviews and confirms the record.
1. Integration of Azure Document Intelligence with the Application
Configure Azure Document Intelligence:
Create an Azure Form Recognizer resource in the Azure portal.
Use its “Custom Model” functionality to train the AI with specific invoice examples (at least 5 for each format).
Export the model and get the API credentials (Endpoint and Key).
2. Application Development:
Create a module that allows uploading invoices (PDF, images) from the backend, portal, directory, email, etc .
Use Azure credentials to send documents to the API and receive processed responses.
3. Process the received information:
Map the extracted fields (invoice number, date, total, tax, vendor, etc.) to the prefixed field templates in the application.
Add logic to validate and save this data in the corresponding records.
2. Application Functionalities
Upload invoices: Allow users to upload documents manually or process them automatically from emails or shared folders.
Integration with Azure API: Send the uploaded documents to the Azure API and receive the extracted data.
Visualization and editing: Display the extracted data in an editable form before confirming the registration in the application.
Additional training: A simple system within the application to add new invoice examples that the AI does not recognize well and retrain the model in Azure.
3. Application Flow
1. User loads or selects an invoice.
2. Application sends the invoice to Azure Document Intelligence.
3. Azure returns the extracted data.
4. The data is automatically mapped into an invoice form in the Application.
5. The user reviews and confirms the record.