AiMetaOff / GUIDES

Inspect metadata in a Canva AI design export

Check what your Canva design export carries and how export format and source assets affect the fields you can inspect.

A Canva flyer can contain a camera photo, a Magic Media illustration and a text headline on the same page. When you export the flyer, the PNG or JPG is the file your recipient gets. Inspect that file to see which metadata it carries, especially if you want to check for a photo’s location data or understand how the design records AI use.

This guide draws on Canva’s documentation. The café flyer is an illustrative example. Authentic Canva PNG and JPG exports were not independently compared.

Generated element versus exported design

In the Canva editor, open Magic Media from the side panel or find it under Apps. On the Images tab, enter a prompt, generate images and select a result to add it to the design. Canva also offers standalone AI image tools that let you download a result or continue editing it. A result downloaded from a standalone tool and a finished page exported from the editor are two different files to inspect. Canva’s Magic Media instructions · Canva’s standalone AI tools

A prompt entered in Magic Media is a generation input. Canva’s public workflow instructions describe how to enter it and place the result, but give no field map for prompt text in a design download. If you need the exact prompt later, keep it with your project notes. To see what a recipient can read from a specific file, inspect the embedded text and XMP entries that the viewer reports.

Canva says it adds AI-generated content metadata for features that create or substantially change content, such as Magic Media and Magic Edit. Its help page says Background Remover alone does not add Canva’s AI-generated content metadata. That describes when Canva applies its AI metadata. The page leaves open which records are written to each PNG or JPG workflow. Canva’s explanation of AI metadata by feature

Compare PNG and JPG from the same Canva page

In the editor, choose Share → Download, select PNG or JPG, choose the page if the design has several, then download. Some size, quality and compression controls depend on the format and plan. Canva also documents a separate download workflow for selected design elements. When comparing formats, export the same page at the same dimensions and keep both direct downloads. If you test an element selection, label that file separately so you can tell which workflow produced it. Canva’s design download instructions · Canva’s selection download instructions

PNG export

Choose PNG when you need lossless image compression or a transparent background, where available in Canva. PNG can store EXIF and text chunks, including XMP. Check the actual download for those records. The format’s ability to hold a text chunk leaves the question of whether Canva wrote a prompt or caption to that file open. Canva’s format guide · W3C PNG specification

JPG export

Canva calls the format JPG. It is JPEG, a lossy format often used for photographs. JPEG can carry EXIF, XMP and IPTC records. Quality settings change how the image is encoded. To see whether the JPG and PNG carry the same fields, inspect each download individually. Canva’s format guide · IPTC photo metadata standard

Inspect the imported photo and the finished design

Imagine a café flyer that combines your own camera photo with a Magic Media illustration. Start with a photo whose EXIF location or XMP description you can identify. Compare it with a PNG and a JPG of a page containing the photo and a headline. Then duplicate the page, add the Magic Media illustration and export the same two formats again. Keep the dimensions and available settings consistent between the two versions of each format.

The first comparison shows whether the source field appears in the finished design. The second helps you see what changes when you add a generated element. Cropping, resizing, compositing or converting an image produces a derivative whose metadata you should check separately. The Canva export instructions describe the available formats and controls. Whether the original camera tags or XMP values carry through remains a question for those specific files. See how to read EXIF, IPTC and XMP fields when comparing the reports.

Check what the downloaded file retains

Open the Metadata Viewer and drop in the direct Canva download. Inspection runs on your device.
Open the file’s metadata report. Compare supported EXIF, XMP, IPTC and text entries where available. If you are checking for a prompt, look for the actual words in the reported values. If you are checking a photo’s location, compare the reported GPS fields with the original.
Review the Content Credentials result separately. If the viewer can read an embedded credential, check the declared source, signature and content binding. Read any unsupported or incomplete check as a limit of the report.

Inspect a Canva design export

Does Canva add C2PA Content Credentials?

Canva documents AI-generated content metadata for certain features. Its AI Product Terms also name C2PA as an example of provenance metadata. The reviewed documentation gives no PNG-by-JPG breakdown of which Canva exports contain a signed C2PA manifest. A Content Credential is a signed record of provenance bound to an asset. Its presence is evidence about the claims it contains. A file with no readable credential leaves its origin undetermined. Canva’s AI Product Terms · C2PA’s Content Credentials explainer

Remove supported metadata from a photo export and verify the copy

Canva’s AI Product Terms prohibit removing, altering or disabling provenance or metadata tags on AI-generated content, including C2PA tags. Use the following steps for a Canva design made with your own non-AI photo and text, where you have permission to remove the selected records. Canva’s AI Product Terms

Inspect the exported photo design and identify the category you want to clean. For example, if the export contains GPS coordinates in EXIF, review the EXIF cleanup option. If the report shows a category is unsupported for that file, leave it out of the cleanup.
In Cleanup scope, select only the supported categories you intend to remove. Cleanup can affect the whole category, including other fields you might want to keep. Select the file, choose Clean selected and download the resulting copy.
Add the downloaded copy to a fresh Metadata Viewer session. Compare the relevant fields with the report from the Canva download. For a GPS check, look for the coordinates themselves because EXIF orientation may remain. Open the copy and confirm that the design still displays as expected. For a second check, inspect the downloaded copy with an independent reader such as ExifTool.

A caption printed on the flyer is part of the rendered image. AiMetaOff works on supported embedded records, so the visible caption and any watermark encoded in the pixels are outside this cleanup scope. See how to verify metadata removal for more on comparing the two reports.

Will metadata removal affect AI labels?

A social platform applies its own label to a post. Meta describes using industry-standard AI indicators and creator disclosures for its AI info label. A local file report cannot predict that decision, and removing a supported embedded record cannot guarantee a change to a platform label. Follow the platform’s disclosure rules when sharing AI-generated content. Meta’s explanation of AI info labels

FAQ

Can I recover a Magic Media prompt from a Canva PNG?

Only if the prompt is present in a readable record or is visible in the image itself. Canva documents the prompt as input to Magic Media and gives no general promise that it is stored in the PNG. Check the reported text and XMP values for the exact prompt.

Does Canva always remove GPS or camera EXIF from uploaded photos?

Canva’s download instructions give no blanket answer. Inspect the uploaded original and the exported page. If the original contains GPS, look for the same values in the export you plan to share.

Does exporting as JPG instead of PNG remove Content Credentials?

Both containers can carry C2PA manifests. Switching formats gives you a new file to inspect. Check its credential result directly. C2PA specification

Can I use the same cleanup steps for a Canva PDF?

PDF has a different cleanup scope. Document properties and document-level XMP are distinct from metadata inside an embedded photo. Use a PDF-specific inspection workflow when the deliverable is a document.