Machine learning for invoice extraction.
***
While being grateful for your interest in the service, the developer of this service moved on to a new project to solve the world peace problem. No new development is planned for this service.
Thank you!
- 2019.11
***
MLReader provides a free web service to enable automatic invoice processing.
The system analyzes an uploaded PDF or image file, and returns key information in seconds.
-
Vendor Name
-
Vendor Address
-
Invoice Number
-
Invoice Date
-
Due Amount
-
Due Date
In countries like the United States, there is no unified invoice. Businesses can design their invoices in any way. These invoices have different sizes, forms, fonts, colors, abbreviations, etc... Although humans magically make sense out of them, it is still a challenge for a computer to decipher them. As the name implies, MLReader utilizes advanced machine learning techniques to automate your business processing.
Access & Data Security
This service is open to anyone who wants to give it a try. There is no signup, and no fee. MLReader is hosted on the Google Cloud Platform, in its US central region. It is a container-based deployment with ephemeral storage. We also utilize Google Vision API to perform optical character recognition. Additionally, we do not save your content or the extraction unless you turn on the 'donate' flag while calling the API.
Please take this into consideration for your regulatory compliance. We do not claim any security clearance, so please don't send us the invoice you got from the CIA.
Accuracy
It depends. There are two major steps in this process: OCR (converting image character to text), and model evaluation. We use Google Vision API for the OCR conversion. It is state-of-the-art technology in this field. However, it also depends on the input file. For example, this awesome invoice file will produce a more accurate result if compared to this low quality file. The quality of the OCR process sets the ceiling for the accuracy of the model evaluation. The model is pre-trained offline. This means that the quality depends on the training dataset. Therefore, we would be grateful if you choose to donate your invoice to periodically retrain the model.
Let's Make Things Better
This service is meant to demonstrate the potential of utilizing advanced ML models in business processing. It is for learning and testing. Although there is no guarantee to match the requirement of your production service, we are happy to work with you to address your specific needs. If you have any of the following concerns, please contact us via the form below or email 'info@mlreader.com'.
-
Want to deploy a separate service with private access from your private cloud
-
Want to have a dedicated model trained with your own dataset
-
Need to extract additional invoice fields from a document
-
Need to extract information from other type of document
-
Need to train the model to understand languages other than US English
-
Need high availability, scalability, or any SLA (service level agreement)
-
Any other question pertinent to this service
Leave a message, or try the free API
If you have any questions or encounter any problems while testing, please don't hesitate to leave a message using the form below. Thank you!
