How NameQuick works
NameQuick reads what is in each file and names it from that. This page covers what it reads in each file type, the four steps from file to name, and what is sent where.
What NameQuick reads in each file type
What NameQuick can read depends on the file type and on how you run the AI. See every supported format
Documents
PDF, DOCX, PPTX, Pages, Keynote, TXT, MD
- Document type and title (invoice, contract, receipt, report, letter, resume)
- Parties and entities (vendor, client, sender, signatories)
- Key dates and reference numbers (invoice date, due date, PO, account, case)
- Amounts, line items, and totals (subtotal, tax, balance, currency)
- Status and version (draft, final, signed, paid)
- Scanned and image-only PDFs: the AI model reads the text from the page image
- Only the first 15 pages of a longer PDF are read
- Pages, Numbers and Keynote files are read from the preview saved inside the file, not their full text
Images
JPG, PNG, WEBP, HEIC
- Subjects, scenes, and settings (beach, city, wedding, a product shot)
- Landmarks and places (Eiffel Tower, Golden Gate Bridge)
- Activities, animals, food, brands, and logos
- Text inside the image, read by the AI model from the image (signs, labels, slides, handwriting)
- Screenshots and photographed documents (receipts, menus, business cards)
- A date, only if the photo shows one
- Naming does not use the date and place the camera saved in the file. Rules can use them, for example to file photos by the date taken
Data Files
CSV, XLSX
- Column headers and sheet names
- Data patterns and categories
- Date ranges in data
- Excel files are read as a sample, the first rows of each sheet, so totals across all rows are not guaranteed
Audio Files
MP3, WAV, M4A
- What the AI model can tell from the recording
- Needs Managed, or your own Gemini or Vertex AI key. With other providers and local models the file is not named
- No built-in transcription or speaker identification
Video
MP4, MOV
- What the AI model can tell from the video
- Needs Managed, or your own Gemini or Vertex AI key. With other providers and local models the file is not named
- No built-in transcription or speaker identification
From file to name
Every file goes through the same four steps.
Read the file
Files come from a folder you added (new files as they arrive, or files you select in All Files), from Quick Rename, from the rename shortcut, from Send To in Windows Explorer, or as PDF attachments through Email Ingest. NameQuick prepares each file for the AI: the text of Word and text files, the PDF itself, and the image for photos and scans. There is no separate OCR step for naming: the AI model reads scans from the page image.
Understand it with AI
The AI works out what the file is (an invoice, a receipt, a contract, a vacation photo). It runs where you choose: Managed, your own API key (Self-Managed), or a local model on your Mac or Windows PC.
Make a clear name
With Smart Rename the AI picks a descriptive name on its own. With a Preset (Written instructions or a Filename template) you set the pattern, and the AI fills it in from the content.
Apply the name
In a new install, each name is applied as soon as it is ready. To check names first, set Review first in Settings › Renaming, for files you add and for watched folders. Files you add then wait in Quick Rename; files from watched folders wait in All Files, marked with a new name. Each rename is recorded in History as a receipt, and Undo puts the old name back. That can fail, for example if the file has since been moved or another file now has the old name. Undo does not reverse what a Rule did.
What is sent, and where
That depends on how you run the AI.
Managed or your own AI
With Managed, NameQuick sends the file to an AI model on servers in the EU, and your files are not used for AI training. With Self-Managed you use your own key for OpenAI, Anthropic, Gemini, OpenRouter, Mistral, or your own OpenAI-compatible server (Custom), or a local model with Ollama, LM Studio, or MLX. MLX needs an Apple Silicon Mac with macOS 14 or later. On Windows there is also the Built-in local engine, and the Mac download also offers Vertex AI with your own Google Cloud project.
Only what is needed is sent
To name a file, NameQuick sends the prepared material: the text of Word and text files, the PDF itself (at most the first 15 pages), the preview of Pages, Numbers, and Keynote files, and the image for photos and scans. With Managed it is stored only temporarily and deleted after processing (within 24 hours at the latest). With your own key it goes directly to your provider. With a local model the file content stays on your computer; NameQuick still goes online for license checks, model downloads, update checks, and crash reports and usage statistics.
You decide when names are applied
With Review first you see every name before it is applied, and in History you can undo recorded renames. NameQuick does not change what is inside your files. For files arriving in watched folders or through Email Ingest, Rules can move, tag, comment, color-label, archive, or trash them, and change their dates.
Try it on your own files
Download NameQuick and rename your first files with Smart Rename.