HEIF MIME Type

What does image/heif mean?

The image/heif media type identifies a High Efficiency Image File Format container with one or more image items. Unlike a codec-specific media type, it does not promise that those images use HEVC, AV1, JPEG, or another particular coding format.¹

That separation is HEIF's defining idea. The container describes images, sequences, metadata, transformations, and relationships, while the coded image data can come from different codecs. HEIF is built on the ISO Base Media File Format, the same box-based family used by MP4.²

A HEIF file can contain much more than one flattened picture:

• A primary image and additional full-resolution images.
• Thumbnails, tiles, bursts, or timed image sequences.
• Derived images such as grids and overlays.
• Alpha, depth, or other auxiliary images.
• EXIF, XMP, color, HDR, and other metadata.²

How can you identify a HEIF file?

IANA assigns .heif to image/heif and also allows .hif. It records no fixed magic number.¹ A typical file has an ftyp File Type Box near the beginning, but the presence of ftyp alone is not enough. MP4, AVIF, HEIC, and other ISO Base Media File Format files use the same box structure.

For the image/heif media type, IANA requires the file to conform to the mif1 brand and list mif1 as a compatible brand.¹ A reliable detector therefore parses the File Type Box and, when it needs to know whether the pixels are decodable, inspects the image item types and codec configuration too.

Optional codecs, profiles, and itemtypes media-type parameters can describe those requirements more precisely. They are useful when a receiver needs to decide whether it can decode a file before downloading the full payload.¹

Renaming a JPEG or HEIC file to .heif does not change its brands or coded images. Use a format-aware tool such as heif-info, ffprobe, or MP4Box when the extension and the contents disagree.

Where did HEIF come from?

Version 1 of HEIF became an ISO standard in June 2015. The format is standardized as ISO/IEC 23008-12, part of MPEG-H.² IANA registered image/heif and the related image/heic subtype on July 30, 2018.¹

HEIF brought container techniques familiar from video workflows to still images. Coded and derived images are stored as items, while bursts and animations can use timed tracks. The format can represent multiple images, alternate representations, tiles, auxiliary images, and rich metadata without requiring one codec for every use.²

Apple brought HEIF into a large consumer photo ecosystem with iOS 11 and macOS High Sierra in 2017. Apple's usual camera files use HEVC-coded HEIF and commonly carry the .heic extension rather than .heif.³

How to open an image/heif file

Apple Photos and Preview can open the HEIF media supported by current Apple platforms. Apple documents built-in HEIF support in iOS 11 and later, iPadOS, visionOS, and macOS High Sierra 10.13 and later.³

Web support needs more care. Safari 17 added HEIC support, meaning the familiar HEVC-coded part of the HEIF family. Apple recommends offering it as an alternative because support on other browsers and platforms remains limited.⁴ A generic image/heif file may use a different codec, so neither the extension nor the container name alone guarantees that Safari or another application can render it.

For a public website, provide a fallback that browsers can decode broadly:

<picture>
  <source srcset="photo.heif" type="image/heif">
  <source srcset="photo.avif" type="image/avif">
  <img src="photo.jpg" alt="A descriptive alternative">
</picture>

Send an accurate Content-Type and keep a JPEG or PNG fallback. If you know the codec-specific subtype, use it. An HEVC-coded HEIC file belongs under image/heic, while a conforming AV1 Image File Format file belongs under image/avif.

HEIF support in Poly

Poly defines image/heif as an image type and infers it from .heif filenames. It also handles image/heic and .heic separately. IANA's .hif extension is not mapped by Poly's filename-only inference, so a .hif upload needs an accurate media type to enter the HEIF path.

Poly's loader passes both HEIF and HEIC files to libheif, selects the primary image, applies container transformations, and decodes it to RGBA pixels. The normal image pipeline then creates WebP thumbnails, extracts dimensions and readable EXIF fields, and produces a visual embedding for similarity and color search. A repository integration test includes a .heif fixture and verifies dimensions, thumbnails, and an image embedding.

That support has two important boundaries. First, libheif must have a decoder for the codec used by the primary item. The broad image/heif label does not guarantee this. Second, Poly renders the primary image rather than exposing every image, sequence, depth map, metadata item, or relationship in the container.

The viewer uses a browser-friendly rendering or a generated thumbnail, so it does not rely on native browser support for the original file. The Poly agent receives the rendered image and can reason about its visual content. Plain images are not OCR'd, so visible words do not become full-text search terms.

HEIF, HEIC, and AVIF are not synonyms

The three terms describe related but different promises:

An application that supports one member of the family may not support the others. HEVC and AV1 require different decoders, and a generic HEIF file may use another codec. Choose the most specific correct media type instead of using image/heif as a catch-all label for every box-based image file.

Converting HEIF without losing important data

Choose the output based on what needs to survive:

• Use JPEG for broad compatibility when one flattened, opaque image is enough.
• Use PNG for a lossless rendered image or ordinary alpha transparency.
• Use AVIF for modern web delivery after testing browser, HDR, and color requirements.
• Keep the original HEIF when it contains multiple images, edits, depth, HDR, or metadata that a single export cannot represent.

Apple Photos and Preview can export supported HEIF images to JPEG or PNG.³ A conversion can preserve the appearance of the primary image while discarding other items and relationships. It can also flatten HDR into SDR or add another lossy encoding generation.