File Types

Document AI supports the following image types.

For information about file size and page limits, refer to theQuotas andLimits page.

Note: Document AI includes some supported file types inPreview.These will be charged when they are released to General Availability (GA).

Name	File Extension(s)	MIME Type
Portable Document Format (PDF)	.pdf	application/pdf
Graphics Interchange Format (GIF)	.gif	image/gif
Tag Image File Format (TIFF)	.tiff,.tif	image/tiff
Joint Photographic Experts Group (JPEG)	.jpg,.jpeg	image/jpeg
Portable Network Graphics (PNG)	.png	image/png
Bitmap (BMP)	.bmp	image/bmp
WebP	.webp	image/webp
HyperText Markup Language (HTML)	.html	text/html
Microsoft Word Office Open XML (OOXML)	.docx	application/vnd.openxmlformats-officedocument.wordprocessingml.document
Microsoft PowerPoint OOXML	.pptx	application/vnd.openxmlformats-officedocument.presentationml.presentation
Microsoft Excel OOXML	.xlsx	application/vnd.openxmlformats-officedocument.spreadsheetml.sheet

Note that some of these image formats are "lossy" (for example, JPEG). Reducingfile sizes for lossy formats may result in a degradation of image quality and accuracy of results from Document AI.

Note: Prior JPEG compressions for TIFF are unsupported. Type of JPEG encapsulationdefined by the TIFFversion 6.0specification.Note: HTML and OOXML support are only available withlayoutparser.Customsplitter only supports PDF, TIFF, TIF, andGIF file types.

Document scan resolution

For most accurate OCR results from Document AI, document scans should bea minimum of 200 dpi(dots per inch).300 dpi and higher generally produce the best results. OCR accuracy is dependenton both the resolution and the minimum font size, along with other factors likedocument (and if handwritten, handwriting) quality, so testing is recommended.Theimage quality analysisfeature can help evaluate resolution concerns.

NOTE: 2k x 3k pixels are required for the US driver's license back side image inorder to read the barcode.