Rename PDFs by content, in bulk
Rename PDFs by Content gives a folder of badly named files — scan_0001.pdf, IMG_2041.jpg, download (7).pdf — names that say what they are. Drop in a batch of PDFs, scans or phone photos of documents; the tool reads the first pages of each, AI picks out the document type, date, issuer and reference number, and every file gets a name from a pattern you choose, such as 2026-03-05_Invoice_Sharma-Office-Supplies_INV-2026-0042.pdf.
It suits anyone who files paperwork digitally: invoices and receipts for an accountant, bank statements and payslips for a loan or visa application, insurance and medical letters for the family archive. Dates are written as YYYY-MM-DD, so renamed files sort in date order in any folder.
Everything is shown in an editable table before anything is downloaded, with the AI’s confidence for each value. The files themselves never leave your device.
How to rename PDFs by their content
- Drop your PDFs and images on the upload box. Each file’s first pages are read in your browser; scans and photos are read with OCR.
- Check the name pattern. The default is
{date}_{type}_{issuer}_{reference}; click the field buttons to build your own, and choose how spaces and long names are handled. - Press Read the documents with AI. The table fills with the type, date, issuer and reference of each file and the new name.
- Review anything marked check. Fix a field and the name updates, or type over the new name directly. If dates like 05/03/2026 appear, choose day-first or month-first.
- Press Download renamed files (ZIP). The ZIP holds your original files under their new names, plus a CSV log of old and new names.
Rename PDFs by Content features
Names from what is inside
Type, date, issuer, reference and title come from the text of each document, not from file metadata or upload dates.
Values copied, not invented
The date, issuer and reference are spans that appear in the document; the AI only chooses which candidate plays each role.
Your pattern, safe names
Any order of fields and separators; missing values drop out cleanly, and characters that break file systems (/ \ : * ? " < > |) are removed.
Confidence on every value
Each field shows its confidence, and names built from a value under 60% or an ambiguous date are flagged “check”.
Edit in place
Change the type from a list, fix a date or issuer, or type a completely different name. Duplicate names get _2, _3 automatically.
Unchanged files plus a log
The renamed files are byte-for-byte copies of the originals, and rename-log.csv records every old and new name with the values used.
Name pattern fields
| Field | What it becomes | Example |
|---|---|---|
| {date} | The document’s own date (issue, statement or letter date), as YYYY-MM-DD when the order is certain | 2026-03-05 |
| {type} | One of 19 types: invoice, receipt, bill, bank statement, payslip, tax document, contract, letter… | Invoice |
| {issuer} | The organisation or person that issued the document, with labels such as “From:” removed | Sharma-Office-Supplies-Pvt-Ltd |
| {reference} | Invoice, statement, policy, account or letter number | INV-2026-0042 |
| {title} | The document’s heading line | Account-Statement |
| {original} / {n} | The old file name without its extension / a running number | scan_0001 / 07 |
About dates
A date such as 5 March 2026, March 5, 2026 or 2026-03-05 is converted to 2026-03-05. A date like 05/03/2026 is ambiguous: it is 5 March in India, the UK and most of the world, but May 3 in the US. Such dates are left as written (05-03-2026) and flagged until you choose Day first or Month first; 13/03/2026 needs no choice because only one reading is valid.
Limits and tips
- Poor scans rename poorly: if OCR misreads the date or the issuer, the name will be wrong. Check flagged rows, and rescan very faint pages.
- Non-English documents work, but less reliably than English.
- Reading the first two pages is usually enough. Choose the first page only for speed, or three pages when the key details sit on a cover sheet’s second page.
- Up to 200 files per batch. Files without readable text keep their original names.
- One PDF containing several documents? Separate it first with Split PDF by Content. To combine renamed files in name order, use Merge PDF by Filename; to pull amounts and tax from invoices, the PDF Invoice Extractor.
Rename PDFs by Content FAQ
How do I rename PDF files based on their content?
Drop the PDFs on Rename PDFs by Content, press “Read the documents with AI”, review the new names in the table and download the ZIP. It is free, needs no account and handles many files at once.
Are my files uploaded?
No. The files stay on your device and the ZIP is built in your browser. Only the extracted text of each file’s first pages is sent, when you press the AI button, and it is not stored.
Can it rename scanned PDFs and photos?
Yes. Scans and JPG, PNG or WebP photos are read with OCR in your browser first. Clear, straight images give the best names.
What if the AI picks the wrong date or issuer?
Edit the value in the table and the name updates, or type a new name directly. Values with low confidence are marked “check” so you know where to look.
Does it change the files themselves?
No. Only the names change; the PDFs and images inside the ZIP are identical to the originals.
Why is my date written as 05-03-2026 instead of 2026-03-05?
The date could be 5 March or May 3. Choose “Day first” or “Month first” under “Dates like 05/03/2026” and the name switches to YYYY-MM-DD.