Interface Overview

This guide gives you a brief walk-through of all the sections in your model.

If you've had some time to explore Nanonets, you're likely already familiar with the main sections, such as My Models. Dive into the Interface Overview for a more detailed understanding of the platform's features and functionalities.

When you click on a model from your My Models list, you will see an Active Model tab on the left-side navigation panel. Explore the sections under Active Model:

AI Training

The AI Training section allows you to improve the accuracy and performance of your model. Here's a breakdown of the sub sections under AI Training:

  1. Manage fields:

    • Add or remove fields that your model should learn to extract from documents.

  2. Training files:

    • Upload sample documents to train your model on.

    • Annotate these documents to provide labeled data for training.

    • Assign files, allowing multiple users to annotate.

  3. Training session:

    • Monitor the training status of your model.

    • View training metrics such as the F1 score to assess performance.

Extract data

The Extract Data section provides you with a structured view of your captured data and offers functionalities for document management. Here's an overview of its key functionalities:

  1. Document Management:

    • View captured data in a custom structured format. (Eg: View invoices by vendor_name or reverse order of due_date)

    • Search and filter files based on criteria like Uploaded by, Uploaded at, Updated at, Verified at, File name, Validation Status, and Review status.

    • Bulk actions include: Assign to member, Change review status, Retry upload, Retry Export and Delete.

    • At the bottom, you will see 2 prominent buttons for key actions: Upload files, Download extracted data.

  1. Document Review and Editing:

    • Click on a file to review individual documents.

    • On the left-hand side, the selected image will be shown (with tools to annotate, zoom and scroll)

    • On the right-hand side, the captured flat fields will be listed first, followed by line items.

    • Under the file name, you can toggle between pages in a multi-page PDF

    • From the three-dots menu, find more advanced actions (like re-run workflow, view file edit history, show/hide fields from view etc.

    • Fields flagged based on predefined rules, will display red error marks. You can hover over these to see what the error is.

    • Use the Approve or Review buttons below to change file status.

Workflow

The Workflow section offers functionality to automate various document processing tasks beyond data capture. Here's a breakdown of the sub sections under Workflow:

  1. Import:

    • Set up automated imports via Email, API, Google Drive or a Nanonets Document Classification model.

  2. Nanonets AI:

    • This is just to indicate which model the imported files are sent to.

  3. Data actions:

    • Add data actions as steps in your workflow to modify and enrich captured data.

    • Select actions such as formatting, lookup from external sources, regex operations, math functions, and more.

  4. Approvals:

    • Set up review stages and approval rules, ensuring thorough assessment and resolution of discrepancies or exceptions.

    • Flag and assign files for review by team members if they do not meet specified conditions.

  5. Export:

    • Automate export of structured results to various applications, such as Quickbooks, Salesforce, Google Sheets and more.

Model settings

The Model Settings page provides an overview of the settings and configuration options for a specific model. Here's a breakdown of the different sections and their functionalities:

  1. Model information:

    • Model ID: Unique identifier assigned to the model

    • Model Description: Brief description or name of the model, for additional context or to your team.

    • Model Plan: Manage the plan associated with your model

  2. File Upload Settings:

    • Sync/Async: Choose between "Sync" for immediate file upload and processing, or "Async" for deferred processing (within 5 minutes), recommended for large file volumes.

    • PDF Page Selection: Select specific pages or page ranges within multi-page PDFs to process with the model.

    • Processing Type: Choose between treating each page as a separate document ("Per page") or treating all pages as one document ("Per document") for processing multi-page files.

  3. Email Import Settings:

    • Process email body if no attachments are found

    • Add email metadata to response: Store sender email, subject name, email timestamp, and email contents as part of the request metadata for access in the API response.

  4. Table Data Capture: Enable or disable tabular data capture from documents.

Last updated