Why Do My Photos Have No Date or Location, and Can I Get It Back?

Why do my photos have no date or location?

Because the platform removed it. Facebook, Instagram, X, WhatsApp, Discord, Slack and iMessage all strip EXIF from images on upload, and most re-encode the file as well. What you download back is a new JPEG that was never on the camera, so the date taken, the GPS coordinates, the camera body and the lens are not "hidden" in it. They are absent. No tool recovers them from that file, because the bytes that held them were discarded on someone else's server.

Two things survive more often than people expect: the pixels, and whatever the platform chose to keep. Everything else depends on how you got the file.

How do I tell a stripped photo from one that never had metadata?

Read the file and look at WHICH fields are present, not whether any are.

Stripped by a platform. Almost no EXIF, but the image dimensions are a

common web size and the JPEG carries the encoder signature of a server-side re-encode rather than a camera. Often a single Software or ColorSpace tag survives alone.

Never had it. Screenshots, exports from design tools and AI-generated

images usually carry a Software tag naming the tool and nothing about a camera, because no camera was involved.

Intact original. Camera make and model, lens, exposure triple, and

DateTimeOriginal all present together. GPS present only if location was on.

You can check any of this without uploading the file anywhere using the image metadata reader, which parses EXIF, IPTC and XMP in your browser. For video, the equivalent fields live in the container rather than in EXIF, and the video metadata reader reads those.

Which fields can I actually get back?

Ranked by how likely you are to recover them.

What you want

Recoverable from the stripped file

Where it still might exist

Date taken	No	Original file, cloud backup, the message you received it in
GPS location	No	Original file; the platform may keep it server-side without showing you
Camera and lens	No	Original file only
Who took it	Never in EXIF anyway	IPTC creator field, if the photographer set one
What is in the picture	Yes, always	The pixels
Roughly when, from content	Sometimes	The pixels: signage, plates, seasons, screens

The line that matters is the last two. The content of a photograph is the one part no upload pipeline strips, which is why attributes derived from the image survive when the header does not.

Can I recover the date from the picture itself?

Sometimes, and you should be honest about the confidence. A visible timestamp overlay, a newspaper, a scoreboard, a phone screen or a dated sign can date a photo precisely. Seasonal cues, construction stages and vehicle models narrow it to a range. None of this is forensic proof, and a court will not take "the trees had no leaves" as a date.

What it IS good for is search. If you are trying to find one photo among hundreds of thousands, a derived approximate date plus what the picture contains is usually enough to get you to the right handful, which is the actual job most of the time.

What if I have to do this across a whole archive?

Reading one file in a browser stops working at about the point you stop doing it by hand. For a library where the metadata is inconsistent, partially stripped or was never written, the durable answer is to stop depending on the header and index what is in the content.

That means running the images through models that produce searchable attributes, then keeping those attributes next to the object rather than in the file:

Objects stay where they are. The photographs stay in your bucket. Nothing

needs re-uploading to a service that will strip the headers again.

Attributes are derived once. Faces, text visible in the image, objects,

scene type and a semantic embedding are computed per image and stored beside it.

The search is over those attributes. "Photos of the loading dock in the

rain" resolves against what the model saw, not against a DateTimeOriginal field that three quarters of the archive is missing.

The trade is worth naming plainly: derived attributes are inferences, and EXIF, when it is present and untampered, is a record. Use the header when you have it. Derive when you do not, and label which one you are looking at.

Mixpeek does this over object storage: you point it at the bucket, it extracts per-image features and keeps them addressable, and you search them together. The models page lists what actually runs, the Mixpeek Vector Store is where the vectors live, and pricing is per feature extracted rather than per seat.

What should I do before uploading, so this does not happen again?

Keep the originals. One archive that nothing re-encodes. Every workflow

that matters reads from there, not from the copy that came back off a chat app.

Decide about GPS deliberately. Stripping location before you publish is

usually the right call for privacy. Stripping it in your own archive is a loss you cannot undo.

Write the fields you care about. IPTC creator, credit and description

survive more pipelines than GPS does, and they are yours to set.

Record provenance outside the file. A checksum and a source note in your

own system outlives any header, because it does not travel with the bytes.

Frequently asked questions

Does every platform strip photo metadata?

Nearly all consumer platforms do on upload, and most re-encode as well. Some keep a subset: a few preserve orientation and colour profile. Cloud storage that syncs a file rather than publishing it (Drive, Dropbox, iCloud in its native flow) usually keeps the original intact, which is why the copy in your backup is often the one that still has the data.

Can a tool undelete EXIF from a photo I downloaded?

No. A tool that claims to is either reading fields that were never removed, or guessing. The stripped fields are not in the file you have.

Is a photo without metadata a sign it was faked?

No, and this is a common mistake. Missing EXIF is the normal state of almost every image on the internet, because uploading removes it. It tells you the file passed through a platform, not that anyone edited the picture.

Why does my screenshot have no camera information?

Because there was no camera. A screenshot is generated by the operating system, so it carries a Software tag and dimensions and nothing about optics. The same is true of exports from design tools and of AI-generated images.

Can I search photos that have no date or location at all?

Yes, and that is the practical answer for most archives. Searching on what the images contain works whether or not the header survived, which is the point of deriving attributes from content.