All Pillars
The Proof Engine

What Really Survives
When You Upload an Image?

You embedded the metadata. You wrote the alt text. You filled in the copyright. Then the platform stripped it all. Here's what actually makes it through — platform by platform, field by field.

The Assumption That Costs You

Most photographers assume metadata survives upload.

Most developers assume alt text comes from the image.

Most marketers assume search engines can "read" their photos.

None of this is reliably true.

The gap between what you embed and what survives is the single biggest reason images fail to perform in search, fail to carry attribution, and fail to be understood by AI systems.

The Metadata Survival Matrix

What happens to your carefully embedded data the moment you hit upload.

Google Images

EXIFPARTIAL
IPTC YES
XMP YES
AltN/A

Reads IPTC/XMP for indexing. Primary discovery channel for embedded metadata.

Facebook

EXIF NO
IPTC NO
XMP NO
AltAUTO-GENERATED

Strips all metadata on upload. Generates its own alt text via ML.

Instagram

EXIF NO
IPTC NO
XMP NO
AltOPTIONAL MANUAL

Complete metadata wipe. Heavy recompression destroys embedded data.

Twitter / X

EXIFPARTIAL
IPTC NO
XMP NO
AltOPTIONAL MANUAL

Strips GPS and most EXIF. IPTC/XMP removed entirely.

WordPress (default)

EXIFPARTIAL
IPTCPARTIAL
XMP NO
AltMANUAL

Resizing can destroy sidecar data. Alt text must be entered in media library.

Squarespace

EXIF NO
IPTC NO
XMP NO
AltMANUAL

Strips and recompresses. No metadata passes through to the rendered page.

Flickr

EXIF YES
IPTC YES
XMP YES
AltN/A

One of the few platforms that preserves and displays full metadata.

Adobe Stock

EXIF YES
IPTC YES
XMPPARTIAL
AltN/A

Preserves IPTC Creator/Copyright. Custom XMP namespaces may be dropped.

The uncomfortable truth:

Only two platforms in this list reliably preserve full IPTC and XMP metadata. The rest strip it partially or entirely — and none of them tell you they did.

The Three Layers — and What Each One Does

Not all metadata is created equal. Understanding the three layers is the first step to understanding what you're losing.

EXIF — The Camera Layer

Written by the camera at the moment of capture. Includes shutter speed, ISO, focal length, GPS coordinates, and date/time.

Survival rate: Moderate. GPS is often stripped for privacy. Technical fields sometimes survive recompression.

IPTC — The Editorial Layer

The standard for professional image metadata. Creator name, copyright notice, caption/description, keywords, and usage rights. This is the layer Google Images actually reads for indexing.

Survival rate: Low to moderate. Most social platforms strip it entirely. Google and Flickr preserve it. WordPress partially retains it.

XMP — The Context Layer

Adobe's extensible metadata framework. Supports structured fields, custom namespaces, and rich context — including licensing, provenance, and semantic descriptions. The most powerful layer, and the most fragile.

Survival rate: Very low. Almost universally stripped on upload. Only specialist platforms preserve it.

What This Means in Practice

Before Upload

  • Creator: John Smith Photography
  • Copyright: © 2025 All rights reserved
  • Description: Aerial view of sustainable architecture…
  • Keywords: architecture, sustainability, green building

After Upload (most platforms)

  • Creator: stripped
  • Copyright: stripped
  • Description: stripped
  • Keywords: stripped

This is why ContextEmbed exists.

Not to fight platforms — they'll strip what they strip. But to ensure that every image leaves your system with the strongest possible metadata profile, embedded to survive wherever the standards allow it.

Stop Hoping. Start Embedding.

ContextEmbed writes structured, standards-compliant metadata into every export — so what can survive, does survive.

Try ContextEmbed Free

3 exports • live web app • no guessing