Auto-Organize Documents into Folders: Complete Guide

Sort a multi-page scan into folders by month or sender automatically: how the ZIP folder option works, what the three layouts look like, and which pages end up in Unknown/.

After a scan run, everything sits in one folder as Scan_001.pdf through Scan_047.pdf, and the March invoice from your internet provider is somewhere in there. Docusplit sorts that out in one upload: it splits the scan into individual documents, names each file by content, and can drop the files into folders by month (2025-03/) or by sender (Acme-Corp/). You pick the layout before you upload; the ZIP comes back already sorted. This guide shows what the three layouts look like, which one fits which job, and what lands in the Unknown/ folder.

Three layouts for the same files

Flat: all files in one folder

downloads/
├── INV-2025-0315_Acme-Corp.pdf
├── INV-2025-0310_Northwind-Ltd.pdf
├── Contract_Blue-Harbor-Dental_2025-02-01.pdf
├── Reminder_City-Utilities_2025-03-20.pdf
└── ...

Fine for a dozen files a month. Since every name starts with an invoice number or a document type, sorting by name already groups the files usefully. At a hundred files, the scrolling starts, and a grouping by folder earns its keep.

By month: one folder per YYYY-MM

downloads/
├── 2025-01/
│   ├── INV-2025-0115_Acme-Corp.pdf
│   └── Contract_Blue-Harbor-Dental_2025-01-02.pdf
├── 2025-02/
│   ├── INV-2025-0210_Northwind-Ltd.pdf
│   └── Notice_IRS_2025-02-28.pdf
├── 2025-03/
│   ├── INV-2025-0315_Acme-Corp.pdf
│   ├── Reminder_City-Utilities_2025-03-20.pdf
│   └── INV-2025-0305_Northwind-Ltd.pdf
└── Unknown/
    └── Document_Page_31.pdf

The month comes from the invoice date or document date the AI reads on the page, not from the day you scanned. That's what makes it useful for bookkeeping: the 2025-03/ folder is the March package, regardless of when the paper reached the scanner. The trade-off is that one vendor's documents are spread across twelve folders per year.

By sender: one folder per company

downloads/
├── Acme-Corp/
│   ├── INV-2025-0115_Acme-Corp.pdf
│   ├── INV-2025-0215_Acme-Corp.pdf
│   └── INV-2025-0315_Acme-Corp.pdf
├── Northwind-Ltd/
│   ├── INV-2025-0210_Northwind-Ltd.pdf
│   └── Contract_Northwind-Ltd_2025-01-15.pdf
├── City-Utilities/
│   └── Reminder_City-Utilities_2025-03-20.pdf
├── IRS/
│   └── Notice_IRS_2025-02-28.pdf
└── Unknown/
    └── Document_Page_31.pdf

Everything from one counterparty in one place: good for vendor files, contract folders, and questions like "what did we get from City Utilities last year". The folder name is the sender as the AI reads it on the page, so "Acme Corp" and "Acme Corporation Inc." can end up as two folders. Chronology gets harder; the ISO dates in the file names bring it back within each folder.

How the folder option works

Choose the mode: invoice mode for a pure invoice batch, document mode for a mixed pile.
Choose the ZIP folder structure on the same screen: no folders, by date, or by sender.
Upload one PDF, up to 50 MB, any number of pages; only your page quota counts.
Download the ZIP. Besides the folders it contains _overview.csv, listing file name, page, number or type, sender and date for every document, and _metadata.json with the same data plus each file's folder path for scripts.

The layout is decided before processing, not at download. To get both views, upload the file a second time with the other option; the second run counts against your quota again. The cheaper route for advanced setups: take the flat ZIP once and let your DMS or a small script build the second grouping from _metadata.json.

Which layout for which job

Monthly bookkeeping. Scan the month's invoices in one batch, invoice mode, folders by date. The month folder goes to your accounting tool or bookkeeper; the CSV is the cover sheet.

Vendor files. Mixed supplier documents (invoices, contracts, delivery notes, reminders), document mode, folders by sender. Drop each company folder into the matching vendor folder in your cloud drive.

Year-end receipts. Run the year's receipts in a few batches with folders by date, then merge the month folders from each ZIP into one tree and hand the accountant twelve folders plus the CSVs instead of a shoe box.

Project files. This is where the feature stops. Folders follow the sender printed on the page, not your project or client names; a quote you sent out carries your own letterhead and would land in your folder, not the client's. Let the tool do the splitting and naming, and do project-based filing in your DMS or by hand afterwards.

What typically goes wrong

A document without a readable date lands in Unknown/ when sorting by month; one without a readable sender does the same when sorting by sender. Sender-name variants split one vendor into two folders. A prominent received stamp can occasionally win over the letter date, putting a document one month off. Pages with nothing readable at all (handwriting, faded thermal paper, scans below roughly 200 dpi) come back as Document_Page_N.pdf with "Not detected" in the CSV. There is no correction step inside the tool: check Unknown/ and the CSV first, then move or rename the few files that need it.

The file names inside the folders

Invoices are named InvoiceNumber_Company.pdf, everything else Type_Sender_Date.pdf, capped at 100 characters, with underscores instead of spaces. That scheme is what makes the folders work: within any folder, files sort by number, type or date without further effort. The full reasoning, plus alternatives if you name files by hand elsewhere, is in How to rename PDF files.

Frequently asked questions

Can I change the folder structure later?

The folder layout is chosen before upload, together with the processing mode. To get a second layout, run the same PDF again with the other option; each run counts against your page quota. Rearranging files inside the extracted ZIP by hand is always possible.

What happens if the date isn't recognized?

With folders by month, a document without a readable date goes into the Unknown/ folder; with folders by sender, a document without a readable sender does the same. Check that folder first and move the files by hand.

Does this work with international documents?

The AI reads documents in any language it can see; German, English, French, Spanish, Italian and Dutch are the ones tested. Folder names are taken from the page itself: the document date for month folders, the company or sender name for sender folders. Document-type labels in the file names come out in English on the English interface.

Can I customize the folder names?

No. Month folders are always YYYY-MM, and sender folders use the company name as the AI reads it, trimmed to 50 characters. Rename folders after download if you need your own scheme.

How are multi-page documents handled?

Continuation pages stay with their first page, so a three-page contract is one file in one folder, not three files spread across folders.

Is this compliant with record retention requirements?

Sorting by month makes files easy to find, which auditors appreciate, but a ZIP on a hard drive is not an audit-ready archive. You still need immutable storage, access logs and a documented process; the rules per country are in Invoice archiving and retention compliance.

Pick one layout and try a batch

Decide whether your filing runs on time (by date) or on relationships (by sender), upload one scan with that option, and see whether the tree matches how you search. The first 30 pages are free after creating an account, no credit card needed. Start for free

Related articles: