Hello everyone, August was a relatively slow month (feature-wise) since many of our developers were on their deserved holiday and hopefully relaxing ????️
So while this release note is lacking in features, it nonetheless packs a punch with regards to new standard document types and fields. This is thanks to a crucial new hire in our product management department, where we now have someone exclusively focused on pushing out new standards. From now on, you should see a steady supply of new standard document types flowing.
If you think we are missing some critical standards, don’t hesitate to reach out! ????
Under the hood, we laid the groundwork for some more prominent features to find their way into the platform in the upcoming weeks. ( I don’t want to spoil too much, but hopefully, you are not too attached to our current user interface or have something against greatly improved extraction quality. ???? )
Step by step, we are getting closer to our goal to become THE platform for document input management (classification & extraction). Our focus was and for the foreseeable future is on
- new standards
- better classification & extraction quality
- and reliability
I’m confident we will get there, but to give you, the people interested in Parashift out there, more insights into our development progress, I’ll start posting these release notes now with a much higher cadency.
If you have any questions at all, feel free to get in touch ????
Cheers ✌️
Thilo
v21.08-1 Release Notes: Features
????️ Added the possibility to (soft-)delete document types and also restore deleted document types. This should give all users with lots and lots of configurations a tool at hand to clean up cluttered configuration lists ????
Standards
In July and August, we were able to win some clients around vehicle processes, for which we added some new document types for all to enjoy now. (papers, purchase contract, insurance policy) ???? We also added Indian identification document types like driver license or identity card. ???? Further, we worked on ever-expanding our list of standard document types by adding, e.g. marriage & death certificates. ⚖️
German & swiss vehicle registration papers
German vehicle papers (“Zulassungsbescheinigung” part I and II, extracting the most essential data for most use cases including registration number, last name, first name and address of vehicle holder, as well as first registration date, manufacturer, model, fuel type capacity, National type approval date and Type (EA 896/897)
Swiss vehicle registration papers
Swiss vehicle papers (“Fahrzeugausweis”), extracting the most important data for most use cases including registration number, address of vehicle holder, as well as first registration date, manufacturer, model and capacity.
Vehicle Purchase Contract
Extraction of buyer and seller, license plate, mileage, vehicle manufacturer, model, color, condition, banking details, document date, contract number, subject, capacity (ccm3) and maximum net power (kWh) and initial registration date, as well as total amount and currency.
Insurance Policy
Extraction of sender (insurance company) and insuree name & address, policy number, policy-type, contract number, contract date and number, net and gross premium, currency, as well as start and end of contract.
Marriage Certificate
Extraction of names and birth dates of both spouses, as well as the wedding day.
Death Certificate
Extraction of first name, last name, birth date and day of death.
Driver License (India)
Extraction of name, address, birth date and drivers license number.
Identity Card (India)
Extraction of name, address, birth date and identity card number.
Simplified Invoice
New simplified invoice with only header data such as sender, receiver, invoice number, date, and total amounts.This makes enhancing this standard document type with more standard fields like line items or banking details, or adding individual fields a lot easier.
Improvements “under the hood”
- made the life of our machine learning engineers easier by providing them better insights into the platform ????
- prepared the (old) user interface for the upcoming design overhaul (focused on “snappiness” and usability ❤️
- deployed some new fancy machine learning stuff which enables us to improve extraction quality of fields like the banking details (iban, bic, bank) or tax ids significantly in one of the following updates ????
Fixes
- fixed some of the most common document upload problems leading to failed documents
- fixed certain documents getting stuck in the process
