A folder full of files called download.pdf and report (1).pdf is easy to create and hard to search. When a browser saves a PDF, the name may come from the server. The date in Finder or File Explorer may be the date you downloaded it. A month later, neither tells you which edition you have or where it came from.
Start with a folder you can test
Create an inbox, a working copy, and a source log. Keep the downloaded originals in the inbox until you have checked the renamed copies. For a practice run, export each row below as a separate one-page PDF. The publishers and documents are fictional. The worked filenames are human-approved targets after looking at the first page, rather than guaranteed model output.
| Original filename | First-page text for the synthetic PDF | Worked filename |
|---|---|---|
download.pdf | Field Notes Lab · Urban Tree Canopy Survey · Published June 14, 2024 · Revision 2 | R001_2024-06-14_Field-Notes-Lab_Urban-Tree-Canopy-Survey_v2.pdf |
report (1).pdf | Coast Lab · Methods for Counting Nesting Sites · © 2023 · No publication date printed | R002_undated_Coast-Lab_Counting-Nesting-Sites.pdf |
document.pdf | City Records Office · Transport Committee Agenda · Published February 3, 2025 · Meeting February 10, 2025 | R003_2025-02-03_City-Records-Office_Transport-Committee-Agenda.pdf |
Capture the source before changing the name
Make one row per PDF in a spreadsheet or a CSV named sources.csv. Assign each file a stable ID such as R001. Record the original filename, the publisher’s landing-page URL, the direct PDF URL when useful, the download date, the stated publication date and its evidence, and the edition or revision. Keep a column for the final filename.
Prefer the publisher’s record or catalog page for the source URL. A direct PDF can be useful when the landing page offers several editions. If you arrived via search, follow the result to the publisher and capture that address. If a PDF has no reliable source, leave the source field blank and record how you obtained it. Use the same ID in the file and the log so the source remains easy to find after a rename.
A starter header is enough for the practice folder.
id,original_filename,final_filename,source_url,pdf_url,published_on,date_evidence,downloaded_on,version,notesKeep publication, event, and download dates separate
Look for an explicit publication date on the title page or publisher’s record. Use YYYY-MM-DD when the whole date is stated and YYYY when you can verify only a publication year. Use undated if there is no reliable publication date. Record any conflicting dates and where you found them.
The survey says when it was published and identifies revision 2. The Coast Lab PDF only shows a copyright year, so the sample keeps the filename undated and records the copyright in the log. The committee agenda gives both a publication date and a meeting date. Use February 3 as the publication date and put February 10 in the notes. Keep the download date in its own column, even when you download every file on the same day.
A pattern for these examples is ID_PublishedDate_Publisher_ShortTitle[_Version].pdf. Preserve the extension, shorten very long titles, and use separators that work on both macOS and Windows. If you need chronological sorting in the folder, put PublishedDate first and keep the stable ID somewhere in the name.
Choose a method and inspect the proposed names
For three files, manual renaming in Finder or File Explorer is quick. Open each PDF, match it to its log row, and type the approved name into the working copy. When a paper already has a reliable bibliographic record, Zotero can rename its attachment using the parent item’s metadata. Check the parent record first, especially the date and title.
A content-based renamer can save typing when PDFs arrive with generic names. I develop RenamerX, one desktop option for macOS and Windows. Its batch view displays original and proposed names and lets you edit or skip suggestions before applying changes. The built-in local model runs on supported hardware after setup. If you choose a cloud provider, check what that provider receives before processing private documents.
Use Batch Rename and review every suggestion before Apply. Assign the stable IDs in the log, then check the suggestions against the PDF first pages. If a suggestion uses 2023 for the undated Coast Lab file, edit it to undated or skip the row. If the agenda receives February 10, correct it to February 3. These examples illustrate review decisions. Test your chosen model on the working copies to see the suggestions it generates. Keep the human check for scans, title pages with multiple dates, and near-duplicate editions.
Verify the result before moving the files
Apply the names to the working copies only. Open one renamed PDF, search for another by its short title, and look up its source using the ID. Compare the number of processed files with the number of completed log rows. If two files would get the same name, inspect their editions before deciding whether to keep both.
Once the sample works, repeat the workflow on a small group of real files. Keep the originals and the source log until you can recover the source by matching the file ID to the log. For sensitive material, use non-sensitive samples while testing and choose a workflow whose data handling you understand.