Auto-renaming PDF after OCR based on content

Hi,

I have been searching for a solution to auto-rename PDFs after OCR based on the content. I don’t want to use rules (some documents might be one-offs). The PDFs are scanned documents like recipes, invoices, confirmations, doctor notes etc.
Does anyone know of a software that does this (either OCR+renaming or just the renaming) and has a workflow around DT to get this done?

thank you

How should the documents actually be named?

with the important entities and the date.
e.g. “DATE-doctor-issue” or “INVOICEDATE-company-product” or “DATE-company-nameOfInvestmentFund-typeOfMessage”

Theoretically this should be doable using few smart rules:

Trying to get the first one working. But I think I am doing something wrong

After applying the rule:

I’ve never seen this before, even in case of documents not containing a date. Did you insert the placeholder via the contextual menu?

i typed it in and wrapped it in %. i cannot find this in the contextual menu BTW

In the contextual menu of the e.g. Change Name field there should be an Insert Placeholder submenu.

there is a submenu but “Sortable Document Date” isn’t among the options

edit: it seems to be the option ‘document date -> 2020-11-12’

Choose the date type, like Oldest Document Date, then choose the date like 2020-12-11.

i got it to work but this is basically the same problem as i have with hazel. every new invoice requires a change in the rules, defeating the purpose of having the rules in the first place.

My example for…

…is able to handle multiple products & companies, see screenshot. But of course it’s hard to tell without knowing the actually desired names and the actual documents and how often you have to process new types of documents.

but you are hardcoding both the product and companies?

Criss’ smart rule covers three distinct variations, i.e., three different kinds of documents. You only entered one criteria while he used an OR method via the pipe (|) to specify multiples.

Yes. A new company/product would require an update of the rule.

well, no. his smart rule doesn’t cover anything as hardcoding entities is the opposite of what you should do. say i buy something from seller ABC, now i need to add ABC to the hardcoded entity list

here is my hazel rule for renaming ‘vanguard statements’


Same principle as Criss’ rule but thats exactly what i would like to avoid

1 Like

I didn’t say it covered everything. I said it covers three types. And yes, of course you’d need to modify the rule to support new variants.

sorry, what 3 types btw?

Documents from Apple, Microsoft, or IBM.