Every ownership dispute is secretly a date dispute. When two people claim the same work, whoever decides between them is really asking one question: who had it first? The strongest answer isn't a better argument. It's a better timestamp, an independent, dated record showing your work existed, in that exact form, before theirs did.

Timestamp evidence for copyright comes in layers, from free and instant to formal and court-grade. A public archive capture costs nothing and takes a minute. File metadata carries creation dates inside the work itself. A SHA-256 hash freezes a specific version at a specific moment. Copyright registration is the government-held version, and the gate on the remedies that pay. One rule governs all of it: capture the evidence before the dispute, not during it.

What a timestamp actually has to prove

A useful timestamp makes two claims at once: that this exact work existed, in this form, on or before a stated date, and that you, not the person now claiming it, controlled it then. Self-made proof collapses on the second claim. Your clock is set by you, your files are edited by you, your word is yours. A record held by someone with no stake in the outcome is worth more than anything on your own drive.

That yields a working hierarchy: a government registration outranks a public archive, an archive you don't control outranks platform logs, and platform logs outrank your own metadata. US copyright exists automatically once a work is fixed in tangible form, so ownership itself is never the live question, provable priority is. A timestamp is one link in that chain; the rest, from creation records to identity, is covered in the full walkthrough of proving content ownership.

Capture their timeline before it disappears

Your dates are half the case; theirs is the other half, and it's the half that evaporates. Infringing pages get edited the moment you complain, deleted after a host warning, or quietly replaced once you go public. Freeze their side the day you find it, before contacting anyone.

The routine is short. Archive-capture the infringing URL with Save Page Now, the next section covers the tool, and save the capture link. Download the infringing files, hash them, and record the digests. Identify the host, if it isn't obvious, you can find out who hosts the site with a quick lookup, and log what the page claims about its own age, since infringers backdate freely. Treat screenshots as notes, not proof: they're trivially editable and carry no independent date. The capture and the hashes are the record.

One boundary to know: the DMCA is US law. Platforms elsewhere run their own notice systems, Canada uses a notice-and-notice regime, and the EU routes takedowns through the Digital Services Act's notice-and-action machinery. The evidence standard doesn't change; the notice mechanics do.

The free standard: archive captures

The same tool freezes your side of the timeline. Drop your published page into the Wayback Machine's Save Page Now the day you publish, and within a minute you hold a dated, third-party copy of your work as it existed then, accepted by hosts, platforms, and courts precisely because you don't control it, and verifiable by anyone from the capture URL. Repeat after any substantial revision: each version is new evidence, and the sequence of dated versions often settles the matter.

Two habits strengthen the layer. Capture underlying assets, not just the page, the image file itself, the hosted PDF, because pages that render content through scripts don't always archive the way you expect. And keep a log of each capture URL and the date you triggered it, so you can produce records on demand.

Know the limits. Archives only reach public pages: member areas, private repositories, and paid course modules behind a login won't capture, and captures are public records by design, so for unpublished or sensitive material, use the hashing route below instead.

The metadata layer inside the work

Metadata carries creation dates inside the work itself, and most media have a native record worth keeping:

  • Photos: RAW files with EXIF, camera model, capture date, sometimes body serial. A RAW file off the original card is close to unarguable; the same date on a JPEG is only suggestive.
  • Video and audio: project files, stems, and camera originals, backed by render logs and upload records.
  • Writing and code: Google Docs version history, git history, CMS revision logs, any system that dates changes and holds them somewhere you don't control.
  • Commerce: course-platform creation dates, distributor release records, sales receipts showing first availability.

The caveat: this layer is self-generated, and self-generated dates bend. EXIF fields edit in seconds. A camera clock can be set to any date before the shot. A git commit carries whatever timestamp the author's machine supplies, so the stronger fact is the platform's arrival record after a push to a public repository. The right way to use EXIF metadata as evidence is as corroboration: original files, original media where it still exists, and independent records that agree on the same dates. One consistent chain beats three loose claims.

Hashing: proving a version existed

A SHA-256 hash is a 64-character fingerprint of a file. Hash the same file twice, same fingerprint; change one comma in a 300-page manuscript and the fingerprint changes completely. That is why a hash recorded with a date proves something no argument can: this exact file, in this exact version, existed at that moment.

The math is the easy part. The deposit is the hard part, because a hash saved in your own notes proves nothing, you wrote the notes. The digest has to land somewhere independent:

  • Publish the hash on a page and archive-capture that page the same day.
  • Commit it to a public repository, where the host's arrival record carries the date.
  • Use a timestamping service, signed timestamps and blockchain-anchored records both exist.

Hashing also covers what archives can't reach: unpublished work. Hash the manuscript before it goes out to publishers, the design files before they reach a client, and note where each was sent. If a copy surfaces later, an identical hash across two files is practically impossible unless one came from the other, coincidence ruled out by arithmetic, not argument.

The technique runs in reverse in our monitoring: each detection is captured and hashed on discovery, freezing the evidence of their copy before it can be deleted. The same math protects your originals.

Registration: the formal version

Registration with the US Copyright Office is the formal version of everything above: a public, government-held, dated record. Under 17 U.S.C. § 411(a), you cannot sue for infringement of a US work until the Office has acted on your application, a 2019 Supreme Court decision confirmed that a suit filed early is dismissed, not paused. Under 17 U.S.C. § 412, statutory damages and attorney's fees require registration before the infringement began, or within three months of first publication. Register late, and even a slam-dunk case recovers only actual damages, often losses you can't document in a content dispute.

Registration also does timestamp work itself. Under 17 U.S.C. § 410(c), a certificate issued before or within five years of publication is prima facie evidence in court of the copyright's validity and the facts stated in the certificate, the government vouching that you claimed this work, on this date.

File within three months of publishing anything that matters. For the money side, how statutory damages work, $750 to $30,000 per work, up to $150,000 for willful infringement, is covered elsewhere, along with when registration is worth it and how the Copyright Claims Board handles smaller disputes. Foreign authors can register too; the sue-first rule applies only to US works, but the remedies gate applies to every work.

Weak timestamps to stop relying on

The poor man's copyright, mailing a copy to yourself and keeping the sealed envelope, is the most durable myth in this field. It fails structurally: a postmark proves a mailing date, not contents, and an envelope can travel unsealed and be filled later. The Copyright Office has long said the technique is no substitute for registration. A self-sent email inherits the defect: date headers are generated by the sender's own server. Keep both if you have them. Never build a case on them.

The subtler trap is over-trusting a platform's upload date. A posted date shows that an account uploaded then; it says nothing about who runs the account, and platforms don't adjudicate between claimants, they compare what each side files and act on the balance. Your job is to make your side of that comparison unarguable: independent records, several layers deep, agreeing on the same dates. That is why the stack matters more than any single timestamp.

Assembling the pack a platform will act on

Takedown notices succeed on paperwork, not eloquence. Under 17 U.S.C. § 512(c)(3), a valid notice must identify the copyrighted work, identify the infringing material specifically enough to locate it, and state, under penalty of perjury, that the information is accurate. Each sworn statement is backed by a timestamp: which work, which version, which date, which page. The filing guide shows where each piece slots into the notice, and the evidence checklist covers the rest of the pack.

Lay both timelines side by side: your creation records, publication captures, and registration; their first appearance, your discovery date, and your captures. When your records predate theirs and the layers agree, most disputes end at the notice stage, hosts remove material on proper notice because their safe harbor protection is conditioned on it.

One caution: a notice is not the last word. The other side can send a counter notice, and unless you file a court action within the statutory window of ten to fourteen business days, the platform restores the material. If the dates are genuinely contested, that window is when the stack you built before the dispute starts to matter in a courtroom instead of a ticket queue.

Frequently asked questions

Does emailing a file to myself prove when I created it?

Not really, the myth costs real cases. An email's date header is generated by the sender's server, and you control both ends of a self-sent message, the same defect that sank the sealed-envelope trick. Your provider's account records add some weight, since they're harder to fake than a header, so keep the email. Treat it as corroboration, never the spine of proof. Registration, an archive capture, or a deposited hash do the job properly.

Are Wayback Machine captures accepted as proof of copyright?

Hosts and platforms accept them routinely, because the archive is independent and every capture is publicly verifiable from its URL. In court, an archived capture is evidence like any other: stronger with clean authentication, the capture link, a record of when you requested it, matching copies of your own, and weaker if the other side produces a credible record predating it. It is corroboration at the top of the free stack, not an unbeatable trump card.

Can EXIF and other file metadata be faked?

Yes, easily, EXIF fields edit in seconds, and a camera clock can be set to any date before the shot. That makes metadata a supporting layer, not a useless one. Original RAW files, files still on original media, and dates corroborated by independent records are hard to argue with, because faking one field leaves the surrounding trail inconsistent. Edited metadata on a copied JPEG, standing alone, proves almost nothing.

How long does copyright registration take, and which date counts?

Certificates routinely take months to arrive, but the wait costs you little. Under 17 U.S.C. § 410(d), the effective date of registration is the day the Copyright Office receives a complete application, deposit, and fee, so a filing made the week you publish carries that week's date. That keeps the three-month window for statutory damages and attorney's fees intact. The delay is in the paperwork reaching you, not in the protection starting.

Can a timestamp prove ownership by itself?

Rarely. Infringement requires copying, and two people can independently create similar works, a complete defense even when the dates run against you. What timestamps do is destroy the other side's story: a work that provably existed in your hands before any contact undermines a claim of independent creation, and an identical hash on their copy is hard to explain away. Dates, proof of access, and similarity are the working combination; a date alone is one third of an argument.

Your next steps

  1. Publish day: run every public page through Save Page Now, hash every unpublished original that matters, and store the digests in two places.
  2. This week: register anything valuable with the US Copyright Office, within three months of publication, or now if that window has passed.
  3. Discovery day: capture the infringing page before contacting anyone, archive capture, downloaded files, hashes, host.
  4. Compare the timelines. If yours predate theirs, file the notice with the pack behind every sworn statement, the DIY route walks through it end to end, and a commercial takedown service is the done-for-you version.
  5. However you file, the timestamp work happens before the dispute, not during it. That timing is the whole trick.