Named by content

Auto-Rename PDF Files
by Content with AI

Last updated: August 2026

The AI reads each page, picks out document type, sender and date, and names the file after them, such as Contract_Northwind-Ltd_2025-01-15.pdf. Invoices get number and company. scan_001.pdf becomes a name you can search for.

Try free View pricing

30 free pagesNo credit cardDeleted after processing

To auto-rename scanned PDF files, Docusplit's AI reads the content of each page, identifies document type, sender and date, and names the file Type_Sender_Date.pdf, or InvoiceNumber_Company.pdf for invoices. One batch PDF becomes separated, consistently named files in a sorted ZIP with a CSV overview. You do not set up patterns or rules.

Renaming by hand means opening, reading and typing, file after file

Your scanner hands out numbers, not names. Finding the file again is still on you.

scan_001.pdf everywhere

The scanner just counts up. Which of the 99 files is the March late notice, you only know after opening it.

Three colleagues, three schemes

invoice_new.pdf, final_v2.pdf and document(1).pdf sit side by side in the folder. Search finds none of them, because nothing in the name says what is inside.

A minute per document, on a good day

Assume opening, reading and typing take a minute per document. A stack of 50 documents then costs close to an hour, and somewhere along the way a typo slips in.

How scan_001.pdf becomes a filename with meaning

Three steps, one ZIP of named files

Upload your PDF

Drop in the PDF from your scanner or scanning app and pick the Documents mode for mixed stacks. If the stack is nothing but invoices, use the Invoices mode.

AI reads each page

The AI first checks if the page is an invoice, then reads the invoice number and company or the document type, sender and date.

Download named files

Each file is named Type_Sender_Date.pdf, invoices InvoiceNumber_Company.pdf. A CSV overview sits in the ZIP next to them.

One naming pattern for the whole folder, with no typing

What changes when names come from the content

Findable by search

With sender and date in the filename, Spotlight or Windows search is enough, years later too. You do not need a DMS for that.

One pattern, no coordination

Whoever uploads, the name follows the same Type_Sender_Date.pdf pattern. The ISO date keeps the folder in chronological order on its own.

From a single receipt to a full binder

The quota counts pages rather than files, so 30 pages are free once and paid plans cover 100, 300 or 1,000 pages per month. How many PDFs you upload within that is up to you.

A filename has to say what is inside, or you end up opening it

In most offices, scan_001.pdf, Invoice_new_final(2).pdf and Document.pdf sit side by side. When three colleagues use three schemes, search does not find the March late notice, and at year-end the accountant asks for receipts that are somewhere in the folder.

One consistent pattern fixes that without new software. With type, sender and date in the name, your operating system's search is enough, and the folder sorts itself by the ISO date. You can still introduce a DMS later; the files will already fit.

GDPR does not prescribe filenames, and neither do record-keeping rules. Those are about retention periods, integrity and documented procedures, which stay the job of your archive and your company. A clear filename makes the filing easier to follow; it does not replace an archive.

Operating system search is enough

One pattern for the whole team

No naming rule in GDPR

The AI reads the page the way you would and builds the name from it

The AI sees each page as an image (GPT-4.1 Vision via the OpenAI API) and reads letterhead, subject line and date the way you would. This is not classic OCR with a text layer; the output PDFs stay unchanged.

In Documents mode it first checks if the page is an invoice. Invoices are named InvoiceNumber_Company.pdf, for example INV-2025-0281_Acme-Corp.pdf. Everything else is named Type_Sender_Date.pdf, for example Lease_Blue-Harbor-Properties_2025-03-15.pdf. The ISO date keeps the folder in chronological order. Amounts, tax or bank details are not read; the tool is not built for that.

Continuation pages are recognized from the content, such as 'Page 2 of 3' or a repeated invoice number, and kept with the first page. At the end you get a ZIP with the named files and a CSV that lists type, sender, date and a confidence value for each of them.

Invoices by number and company

Other documents by type, sender and date

ISO dates for chronological sorting

From paper stack to filed documents in one pass

A stack of letters, invoices and contracts goes through the feeder, and the scanner delivers scan_batch_001.pdf with 30 pages. Uploaded in Documents mode, the AI detects from the content where a new document starts, keeps the three-page contract together and names the output, say INV-2025-0310_City-Utilities.pdf, Lease_Northwind-Ltd_2025-01-15.pdf and Notice_IRS_2025-03-20.pdf.

At roughly 10 seconds per page in Documents mode, the 30-page stack takes around five minutes, and the progress is shown while it runs. The ZIP holds the individual files, optionally in folders by month or by sender, plus the CSV overview. You copy them into your filing system and the batch file can go.

→ Split a scan PDF into individual documents

Upload the batch scan in Documents mode

About 10 seconds per page, progress visible

ZIP with individual files and CSV

What the AI does differently from renaming by hand

By hand means open the file, read it, type a name, next file. Assume that takes a minute per document, and a stack of 50 documents costs close to an hour, with abbreviations drifting somewhere after the twentieth file.

The AI takes about 10 seconds per page in Documents mode, so a few minutes for the same stack, and it does not drift, because each name follows the same pattern. Mistakes still happen. If it cannot read a page, the file is named Document_Page_7.pdf; if it attaches a page without a letterhead to the wrong document, the CSV shows it through page number and confidence.

So you keep control without touching each page. The CSV lists what the AI read, and you check the rows marked 'Not detected' or with low confidence.

A minute per document by hand (assumption)

About 10 seconds per page with AI

CSV shows which files to check

What the AI reads from a scanned page

There are no rules or templates to maintain; the name follows the content of the page.

Document type detected by AI, such as invoice, contract, letter or tax assessment
Sender, company or agency read from the letterhead
Date in ISO format YYYY-MM-DD so the folder sorts chronologically
Invoices named by invoice number and company
Multi-page documents stay together, continuation pages are recognized from content
Plain PDFs that any DMS can import
Optional folders in the ZIP by month or by sender
CSV overview with type, sender, date and confidence per file

Before:

scan_003.pdf

After:

Contract_Northwind-Ltd_2025-01-15.pdf

Where content-based filenames make the difference

Three scenes from office life

Accounting firm, morning mail

Yesterday's stack goes through the feeder. Instead of scan_001 to scan_040, the folder now holds Notice_IRS_2025-03-12.pdf and Letter_Superior-Court_2025-03-11.pdf.

Year-end receipts binder

The year's invoices go through the scanner once and come back as InvoiceNumber_Company.pdf, one file per receipt, the way QuickBooks, Xero or your accountant want them.

Tenant file at a property manager

Lease, move-in inspection report and service charge statement from one scan land in the file as three named PDFs with dates.

Frequently asked about renaming

What naming pattern does Docusplit use?▼

Invoices become InvoiceNumber_Company.pdf, for example INV-2025-0281_Acme-Corp.pdf. Everything else becomes Type_Sender_Date.pdf, for example Contract_Northwind-Ltd_2025-01-15.pdf. The date is in ISO format so files sort chronologically. Custom patterns cannot be configured.

Can I auto rename PDF files for free?▼

Yes. The free tier covers 30 pages once, with a free account and no credit card. Upload the PDF, wait for processing, download the ZIP.

What happens if a document type is not recognized?▼

If the AI cannot classify a page, the file is named Document_Sender_Date.pdf; if it can read nothing at all, the file becomes Document_Page_7.pdf and the CSV says 'Not detected'. Handwriting, very faint thermal receipts, scans below roughly 200 dpi and attachments without a letterhead are the usual causes, and attachments can also end up with the wrong document. Check those files and the confidence column first, then rename by hand where needed.

Does Docusplit also split multi-page PDFs?▼

Yes, splitting and naming happen in the same pass. Upload a batch scan; the AI detects from the content where a new document starts, keeps continuation pages together and names each part.

Does it work with scanned paper documents?▼

Yes. The AI reads each scanned page as an image (GPT-4.1 Vision via the OpenAI API) rather than through a classic OCR text layer. Machine-printed pages from about 200 dpi read reliably, in grayscale or color, in the languages we have tested (English, German, French, Spanish, Italian, Dutch); handwriting and very faint receipts often do not. The output PDFs are copied unchanged, without a text layer.

How is this different from the Docusplit blog guide on renaming?▼

This page describes the tool. The blog guide compares five ways to rename PDFs, from built-in tools and batch utilities to scripts and AI, and explains when each makes sense. Use this page to try the tool and the guide to weigh the options.

Related uses

The neighboring pages to renaming, from splitting to invoice intake.

Split & Rename PDF

Upload one batch scan of mixed invoices, letters and receipts. The AI splits it into separate documents and renames each file by content. Sorted ZIP with CSV.

Learn more

Detect invoices in mixed PDFs

Multiple invoices in one PDF? The AI finds each invoice boundary in the batch, splits it, and names every file by invoice number and company. 30 pages free.