Supported File Formats
Complete list of file formats NameQuick can process, including images, documents, audio, and video.
NameQuick 3 accepts more than 70 file extensions, from PDFs and Office files to camera RAW, audio and video, the same on Mac and Windows. It names each file from what is inside it, not from its old name or extension. Whether a file can be read depends on its format and on how you run the AI. Files with other extensions aren't accepted.
Recognized file formats#
This is the complete list. The last column says what NameQuick sends with Managed, NameQuick's own AI service. The limits are below the table.
| Category | Extensions | What Managed sends |
|---|---|---|
| The PDF itself. For more than 15 pages, only the first 15. | ||
| Images | JPG, JPEG, PNG, GIF, WebP, BMP, TIF, TIFF, HEIC, HEIF | The image as it is. |
| Camera RAW | RAW, ARW, CR2, CR3, NEF, DNG, RAF, ORF, RW2, SR2, PEF, RWL, NRW, SRW, 3FR, IIQ, MOS, KDC, MRW, MEF | The JPEG preview saved inside the file. Without a usable preview, the file can't be read. |
| Vector graphics | SVG, AI, EPS | SVG and PDF-compatible Illustrator files are turned into an image. EPS and older Illustrator files are recognized but can't be read. |
| Word | DOCX | The text of the document. Headers, footers and footnotes aren't read. |
| Excel | XLSX, XLSM | A sample: the first rows of up to 12 sheets. Macros in XLSM files are never run. |
| PowerPoint | PPTX | The text on the slides. Speaker notes aren't read. |
| Apple documents | PAGES, NUMBERS, KEY | The preview saved inside the file (a PDF or an image), not the full text. Without a preview, the file can't be read. |
| Text, markup and code | TXT, MD, CSV, HTML, CSS, JS, PY, XML, RTF, DOC | The text. For files over 256,000 characters, an excerpt from the beginning, middle and end. |
| Audio | MP3, WAV, M4A, AAC, FLAC, OGG, OPUS, AIF, AIFF, AIFC, CAF, WMA | The file as media. What the model makes of it depends on the model. |
| Video | MP4, MPEG, MPG, MOV, AVI, FLV, WEBM, WMV, 3GP | The file as media. What the model makes of it depends on the model. |
A listed extension doesn't mean every AI model can read every file of that type. Test your own files, especially media and formats that rely on a saved preview. What each AI setup can read shows the differences.
How each kind of file is read#
Documents#
NameQuick reads PDFs, Word documents, PowerPoint presentations, spreadsheets and text files so the AI can pick out dates, amounts, vendor names, reference numbers and document types. A PDF with more than 15 pages is named from its first 15 pages, and its receipt in History shows First 15 pages only. Excel files are read as a sample, so totals across all rows aren't guaranteed. Pages, Numbers and Keynote files are read from their saved preview, not their full text.
Scanned PDFs
Images#
The AI model looks at the image itself and can recognize scenes, landmarks, objects, brands and activities. Camera RAW files are read through the JPEG preview saved inside them. For how the date and place saved by the camera are used, see Photo data (EXIF).
Audio#
Audio files are sent to the AI model as media, if your AI setup takes audio. Voice memos, podcasts and music recordings in the formats above are accepted. NameQuick doesn't transcribe audio itself, and it doesn't fall back to naming from the file's details alone: if your setup can't take audio, the file isn't named and NameQuick shows the reason.
Video#
Video files are sent to the AI model as media, if your AI setup takes video. Screen recordings, camera footage and downloads in the formats above are accepted. As with audio, there is no fallback to naming from the file's details alone.
What each AI setup can read#
Caution
Managed reads every category in the table above. With Self-Managed, what reaches the AI depends on the connection:
| Connection | Reads |
|---|---|
| OpenAI, Anthropic, OpenRouter | Text, images and PDFs. No audio or video. |
| Gemini, and Vertex AI (Mac download only) | Everything above, including audio and video. |
| Custom (your own OpenAI-compatible server) | Text, images and PDFs. No audio or video. If the server can't take a PDF file, its pages are sent as images. |
| Local: Ollama, LM Studio, MLX, and Built-in on Windows | Text, images and PDFs. No audio or video. For photos and scans, the model has to be one that reads images. |
The Mac App Store version has no local models. Check that your setup covers the formats you need.
OCR for scanned documents
For naming, NameQuick has no separate built-in OCR step. The AI model reads the text in scanned PDFs and document photos from the page image. With a local model you need one that reads images, and a scanned PDF is read from up to three pages: the first, one from the middle, and the last of its first 15 pages.
How well this works depends on the scan. The model can misread characters or words, so check suggested names for blurry or hard-to-read scans before you apply them. Content Search uses on-device OCR for its local index, but that is separate from naming.
Photo data (EXIF)#
Cameras and phones save details inside each photo, known as EXIF data. Rules read it directly: the date taken, the camera, the image size, and whether GPS data is present. They use it in conditions and in folder paths, for example to file photos by the date taken.
Naming works differently: it doesn't use this saved data. A Preset field such as "date taken" is filled in by the AI from what the image shows, so the value can be missing or wrong. Don't rely on the camera's date or location to decide the suggested name.
Smart Rename names files from the content described on this page, and Content Search finds them by it later.