No. Not by itself.
A timestamp proves a file existed. It does not prove you wrote what is inside it. That gap is the whole game, and it is where every product in this category stops explaining.
So we read every US music copyright decision we could find from 2015 to 2026, 197 of them, looking for a case that turned on a cryptographic proof of creation. There isn't one. Not one of those decisions mentions a blockchain record, a cryptographic timestamp, a timestamping service, or a self-mailed envelope. In eleven years, no US court has been asked.
SongSecure sells a timestamping product. That is an awkward thing for us to publish, and it is why the rest of this article is about what those cases actually turned on, and about the one thing dated documentation is genuinely good at: whoever stole your song does not have the twelve things that came before it.
Here is what closes the gap and what does not, plus what to do this week.
The one-paragraph version
A blockchain timestamp is evidence of possession at a time, and nothing more. It becomes evidence of authorship only when attached to a sequence (the voice memo before the demo, the session file before the stems, the stems before the master) and to a human being who will swear to it. In the US, §411(a) and the Supreme Court's decision in Fourth Estate both address registration as a precondition to suit; both are quoted below. And in the 197 decisions we read, cases were won and lost on access, on whether the shared element was protectable, and on registration paperwork.
Has any US court ruled on a blockchain timestamp?
No. We looked properly, and the absence is the finding.
We searched all 197 decisions for any mention of a blockchain record, a cryptographic timestamp, a trusted timestamping service or a self-mailed envelope, and searched roughly 3,400 further federal opinions the same way. Two people screened the set independently, and their agreement was measured: κ = 0.8297 (n = 203, 95% CI 0.7502–0.9010), which records the reproducibility of the inclusion screen and not of any outcome.
The count is 197 decisions, held in 203 records: six are duplicate entries for a decision already in the file, marked as such in our screening file. 201 of the 203 are retrievable as text; two are not, and we say so rather than round up. 108 of the 197 are merits rulings, across 82 disputes.
| Search term | Hits in the 197 music decisions | Hits in the wider federal corpus |
|---|---|---|
| blockchain, distributed ledger, smart contract, bitcoin, NFT | 0 | 10 files, none a music copyright merits decision |
| trusted timestamp / RFC 3161, and "poor man's copyright" / self-mailed envelope variants | 0 | 0 |
| notarised / notary public | 5 decisions | 77 files, never used to date a musical work |
It does not mean courts have rejected this evidence. No court in our dataset was ever asked. It means the whole category, ours included, is selling something whose courtroom performance in the United States is untested. Anyone telling you their certificate is "court-ready" is describing something no reported US music decision has ever examined.
It also means something more useful: whatever decides these cases, it is not the thing this industry argues about.
What do music copyright cases actually turn on?
If you want the argument for dated proof made by a court rather than a vendor, here it is.
Bowen v. Paisley, M.D. Tenn., 25 August 2016. Amy Elizabeth Connor Bowen, who wrote in Nashville as Lizza Connor, wrote a song called "Remind Me" in 2007–2008. Brad Paisley and Carrie Underwood released a song of the same name in 2011. She sued. Her strongest access evidence was an email with a recording of her song attached. Real, admissible evidence that contained the actual work. The court:
"…she fails to acknowledge that this communication took place in October 2011, well after both the creation and public release of the Paisley Work. The email exchange therefore could not establish access for purposes of copyright infringement."
The only reason that evidence was worthless was its date.
Bowen still lost
She had close to the documentary record a proof-of-creation product dreams of: a dated writing period, a dated demo of herself singing and playing guitar, a dated lyric sheet, a registration effective 3 September 2008, a documented public performance on 3 March 2008 in front of one of the defendants, and a song plugger pitching the work for ten months.
She lost on summary judgment. The only thing the two songs shared was the unprotectable phrase "remind me": "the allegedly infringing Paisley Work is not substantially similar to the plaintiff's copyrighted Work." The court also recorded that "the defendants have presented no evidence to rebut the plaintiff's assertion that she independently created her Work." She won that point. It did not matter.
Documentation answers when and who. It does not answer is what I own protectable or is the other song substantially similar, and those are what most of these cases die on.
What won, and what lost
Batiste v. Lewis, E.D. La. 2019, aff'd 5th Cir. 2020. Macklemore & Ryan Lewis beat a sampling claim on a sworn declaration, but what did the work was what stood behind it: Ryan Lewis had "turned over all of the audio files I had that documented the process of creating the Macklemore & Lewis Songs." Batiste's own theory rested on a Pro Tools session showing an error about 178 missing audio files. The Fifth Circuit: "Lewis testified that this is a 'common message' that appears whenever files are moved across hard drives. It doesn't mean that those files are, in fact, missing." The difference between the two sides is custody and completeness, not file format.
And in We Shall Overcome Foundation v. The Richmond Organization, S.D.N.Y. 2017, the court reconstructed authorship from a handwritten note on a lead sheet, date-stamped 25 October 1960. The best "documentation survived and mattered" example in the dataset, and the technology was a filing cabinet.
Twelve musicians represented themselves. All twelve lost.
We identified every in-scope decision in which the plaintiff or appellant was self-represented: a working musician, no lawyer. There were twelve, and all twelve lost. The failure modes are uniform:
- Access was not plausibly alleged. From one: "Plaintiff fails to allege facts showing that Defendants had access to his song… because he does not cite facts to support a reasonable possibility that Defendants had the opportunity to hear it."
- The shared element was unprotectable: a short phrase, a title, a five-note sequence. "Because the Complaint alleges only that the advertisements are similar to the unprotected five-note sequence in his song, Cates has failed to plausibly allege a copyright infringement."
- Registration was missing, or was for the wrong work.
- Speculation was offered in place of admissible evidence.
Not one of the twelve lost because they could not prove they wrote it first.
The barrier here was not authorship. It was proving contact, and owning something protectable. A proof-of-creation record helps with one element of a claim and does nothing for the elements these twelve lost on. Two caveats: this is a count of what we read, hand-identified by one reader, not a win rate. And claims that settle or are abandoned produce no written decision and are invisible here.
One thing nobody tells self-represented musicians: in Prunty v. Vivendi, having lost, the plaintiff then faced the defendants' motion for attorney's fees, and the court granted it in part.
What is the most common registration mistake musicians make?
If you take one practical thing from this article, take this. It has nothing to do with blockchain.
The US Copyright Office publishes a two-page explainer, Musical Works, Sound Recordings & Copyright (M-309, rev. February 2020). It opens:
"When you record a song, you may be creating two works that can be protected by copyright law: a musical work and a sound recording. These works are subject to different rules under the Copyright Act and are commonly owned and licensed separately."
The same page defines each:
"A musical work is a song's underlying composition created by a songwriter or composer along with any accompanying lyrics."
"A sound recording is a series of musical, spoken, or other sounds fixed in a recording medium, such as a CD or digital file, called a 'phonorecord.' … The author of a sound recording can be the performer who is being recorded, the record producer who processes and fixes the sounds, both, or even another entity if the work qualifies as a work made for hire."
"Even though musical works and sound recordings are separate works under copyright law, you may be able to register them together on a single application… if the copyright owner of the sound recording and musical work are the same."
Three disputes in our dataset were lost or gutted on that distinction: Nwosuocha (twice, district and appeal), Pickett and Stringer. A fourth, Buchanan, lost nine claims to the adjacent failure: no registration at all.
Nwosuocha v. Glover, the "This Is America" case. The plaintiff registered his album in 2017 on Form SR, as a sound recording, then sued over the composition. In a non-precedential summary order (one that binds these parties but cannot be cited as precedent later), the Second Circuit affirmed dismissal in 2024: "We affirm the district court on the issue of Nwosuocha's registration and so do not reach whether Appellees' song is infringing." It added: "But we do not invalidate Nwosuocha's registration. Rather, Nwosuocha's problem is that his copyright registration is simply for the wrong work."
Pickett v. Migos Touring, S.D.N.Y. 2019, the identical trap: "Plaintiff did not obtain a certificate of registration for his musical composition." Stringer v. Richard, N.D. Ohio 2022, conceded in the plaintiffs' own filing: "there was a mistake in law by the Plaintiff in selecting 'Music Composition'." They asked for leave to register the recording properly; the court refused. Buchanan v. Sony Music, D.D.C. 2020: nine claims dismissed as unregistered. What this looks like for a producer whose instrumental was used after they sent it is set out in what to do when a beat you sent comes back on somebody else's release.
A wrong creation date on the certificate
In Roberts v. Gordy, 877 F.3d 1024 (11th Cir. 2017), the "Hustlin'" case, the Eleventh Circuit recorded what was on the paperwork:
"The second copyright registration, PA 1-334-589, incorrectly stated that the creation date was 2006 (instead of 2005) and did not disclose that there was a prior registration."
The district court referred the question to the Register of Copyrights under §411(b)(2), and the Eleventh Circuit recorded the answer: "the Copyright Office advised that it 'would have refused [the second] registration' had it known of the incorrect creation date." The district court invalidated all three registrations and dismissed the case at summary judgment.
The Eleventh Circuit reversed, and stated the test it applied:
"Therefore, in order to invalidate a registration, (1) the application must contain inaccuracies, (2) the inaccuracies must be material, and (3) the applicant must have the required scienter of intentional or purposeful concealment."
Its closing line:
"In short, the Appellants were erroneously 'hustled' out of court, and now deserve to be heard on the merits."
They won the appeal. On remand, one of the three named plaintiffs was denied standing, and the case settled before trial on undisclosed terms. A wrong creation date on the certificate cost these plaintiffs a district court dismissal and a federal appeal.
Five years later, in Unicolors, Inc. v. H&M Hennes & Mauritz, L.P., 595 U.S. 178 (2022), Justice Breyer wrote for the Court:
"In our view, however, §411(b) does not distinguish between a mistake of law and a mistake of fact. Lack of knowledge of either fact or law can excuse an inaccuracy in a copyright registration. We therefore vacate the Court of Appeals' contrary holding."
Justice Thomas, dissenting, named the split the Court had taken the case to resolve, and Roberts as one side of it:
"We took this case to resolve an apparent split between the Eleventh Circuit, which has held that §411(b)(1)(A) requires 'deceptive intent,' Roberts v. Gordy, 877 F. 3d 1024, 1030 (2017), and the Ninth Circuit, which held below that 'there is no such intent-to-defraud requirement'…"
And later in the same dissent:
"The Court does not decide the question that has split the Courts of Appeals, but instead decides a question that no court has addressed."
Those are the documents. We are not going to tell you what they add up to.
The door you cannot open without registration
17 U.S.C. §411(a): "no civil action for infringement of the copyright in any United States work shall be instituted until preregistration or registration of the copyright claim has been made in accordance with this title."
In Fourth Estate Public Benefit Corp. v. Wall-Street.com, LLC (2019), Justice Ginsburg wrote for a unanimous Court:
"We hold, in accord with the United States Court of Appeals for the Eleventh Circuit, that registration occurs, and a copyright claimant may commence an infringement suit, when the Copyright Office registers a copyright. Upon registration of the copyright, however, a copyright owner can recover for infringement that occurred both before and after registration."
The Copyright Office says the same thing in Copyright Basics (Circular 1): "Before an infringement suit may be filed in court, registration (or refusal) is necessary for U.S. works." And: "the Copyright Office must approve or refuse your application before you can file a lawsuit for copyright infringement, except in cases involving a non-U.S. work."
New Day Worldwide v. Taylor Swift, C.D. Cal. 2020, shows it on the ground: alleging only "that Graham applied for copyright registration… is insufficient. This deficiency is fatal and amendment would be futile." (That one sits in our universe but was screened out of the merits set as a threshold ruling.)
What timestamping can honestly do is cover the window before the Register acts.
What to do this week
Nothing here requires a lawyer, a subscription or a blockchain. Do these in order. A shorter version of the same sequence is set out as a 15-minute protection checklist.
1. Start the sequence habit today, on the song you are working on now. Keep the voice memo, the first DAW session file rather than just the bounce, the lyric draft with the crossings-out, the stems, then the master. Do not tidy them and do not overwrite them. That is the Batiste lesson, and it costs nothing.
2. Check what form your last three registrations were filed on. The Copyright Office's own distinction is quoted above: a musical work is "a song's underlying composition… along with any accompanying lyrics"; a sound recording is "a series of musical, spoken, or other sounds fixed in a recording medium." Which form fits your particular work is a question for the Copyright Office's guidance or a lawyer; we are not going to answer that for you. But nobody needs to tell you which form you already used. It is printed on your certificate. Go and look.
3. Register, and register early. Two clocks reward it, and neither cares how good your timestamp is. The Copyright Office states both in Circular 1:
"Before an infringement suit may be filed in court, registration (or refusal) is necessary for U.S. works."
"When registration is made prior to infringement or within three months after publication of a work, a copyright owner is eligible for statutory damages, attorneys' fees, and costs."
4. Write the split sheet at the session, not afterwards. A split sheet is one page: the song and the date, everyone who was in the room, what each contributed, the percentage each takes of the composition, and each writer's performing-rights society and IPI number. Everybody signs. It is filed nowhere and has no official status. It is a record of what people agreed while they still agreed.
If someone will not sign, act while it is fresh: write down what you understood the splits to be, send it by a method that leaves a record ("here's what I have us at, shout if that's wrong") and keep whatever comes back, including nothing. A dated, unanswered message beats two memories. Whether a contribution makes someone a legal co-author is a genuinely contested question we cannot answer, which is the point of a split sheet. It stops you needing the answer.
5. Hash it before you send it to anyone. Including your collaborator. The ledger is a first-to-hash register, not a first-to-create register: whoever records first, records first. That is not a reason to distrust it, it is an instruction about order. The record goes on before the file leaves your machine, not after somebody starts asking questions.
6. Keep a delivery record every time you send a song to anyone. Loomis v. Cornish, 9th Cir. 2016, is the lesson: the court accepted that he had delivered his song to a UMG A&R representative, and he lost anyway, because access has to reach the people who actually wrote the accused song. So the record worth keeping is not "I sent it." It is who, by name and role; what came back, including a rejection; and whether they were anywhere near the writers. Send from an account you will still control in five years, and keep the reply. Which delivery records courts have credited, and which they have rejected, is set out decision by decision in what courts have accepted as evidence that someone had the opportunity to hear a song.
What if it has already happened?
Everything above is about the window before a dispute. If you have already heard your melody on somebody else's release, the order of operations changes, and that is a different page: what to do when someone has already used your song, covering evidence preservation, registering after discovery, the Copyright Claims Board, demand letters and platform takedowns, with the statutes quoted.
If you lease beats, use samples, or read this from outside the US
If you sell or lease beats. 17 U.S.C. §204(a), in full:
"A transfer of copyright ownership, other than by operation of law, is not valid unless an instrument of conveyance, or a note or memorandum of the transfer, is in writing and signed by the owner of the rights conveyed or such owner's duly authorized agent."
Read your marketplace's default contract against that sentence. A timestamp is silent about all of it. What it answers is which of you had the beat first, and in what form.
If your track contains a sample. This is the one place where you are the defendant, and the advice inverts. The Copyright Office, on sound recordings: "Anyone else who wants to use a sound recording must either get a license from the copyright owner, use a statutory license, or have an exemption apply, like fair use." It says the same of musical works. On an uncleared sample, a timestamp does the opposite of protecting you: it is a dated, tamper-evident, third-party-verifiable record that on a specific day you possessed a file containing someone else's work. Every property that makes it good evidence for you makes it good evidence against you.
If you are not in the United States. "Register your work" appears all over this page, and it is US-specific. GOV.UK's guidance How copyright protects your work states: "You get copyright protection automatically – you do not have to apply or pay a fee. There is not a register of copyright works in the UK." Protection on creation without formalities is the position across the countries party to the Berne Convention. §411(a), §412 and the Copyright Claims Board are mechanisms of one country. The Copyright Office does register works by non-US authors; Circular 1 carves out the non-US case expressly, in the sentence quoted above.
Can someone backdate a blockchain timestamp?
Not in the direction that matters. A blockchain timestamp establishes an upper bound: this file existed no later than block N. On Bitcoin, a block's timestamp must exceed the median of the previous eleven blocks and fall below network-adjusted time plus about two hours. The precision is hours, not seconds — but you cannot make a public chain show an earlier date than reality. What the criticism actually describes is hashing an old file today, which makes your timestamp later than your real creation date. Bing's AI answer, captured 7 August 2026, attributes to chainscorelabs.com the line that "it's possible to backdate data before on-chain submission".
So the objection, correctly stated, is not backdating. It is appropriation: someone hashing a file they did not write. If a collaborator hashes your demo before you do, the ledger will faithfully record their priority. That is a real weakness of every timestamping product on the market, ours included. The only answer to it is step 5: hash first, then send.
Does existence prove authorship?
No. Everything connecting a file to a human is off-chain. Perplexity's answer to this query (captured 7 August 2026) puts it well: "Courts assess the link between the timestamped file and the person claiming authorship; a timestamp alone does not establish that link." Two things close the gap, and neither is technology.
First, the sequence. One hash on one finished MP3 is a possession claim. A dated chain of artefacts is an authorship narrative. That is what Batiste demonstrates: the producer who could hand over every file documenting the creation process won; the party pointing at a single software artefact lost.
Second, a human who will swear to it. Federal Rule of Evidence 902 opens: "The following items of evidence are self-authenticating; they require no extrinsic evidence of authenticity in order to be admitted." Every route in it that could plausibly apply to a record like this, 902(11)–(14) and Vermont's 12 V.S.A. §1913, requires a certification by a qualified person. There is no unattended path from a ledger to an exhibit. Somebody signs.
Does any US law actually cover blockchain records?
Some. None of it is a holding that such a record must be admitted, but "US law has no mechanism for this" is wrong.
FRE 902(14), effective 1 December 2017, provides for self-authentication of "Data copied from an electronic device, storage medium, or file, if authenticated by a process of digital identification, as shown by a certification of a qualified person that complies with the certification requirements of Rule 902(11) or (12). The proponent also must meet the notice requirements of Rule 902(11)." The Advisory Committee Note names the mechanism: "identical hash values for the original and copy reliably attest to the fact that they are exact duplicates. This amendment allows self-authentication by a certification of a qualified person that she checked the hash value of the proffered item and that it was identical to the original."
Two limits travel with it, both from the rule-makers. Per the Advisory Committee Note to 902(13): "A certification under this Rule can establish only that the proffered item has satisfied the admissibility requirements for authenticity. The opponent remains free to object to admissibility of the proffered item on other grounds." And the 902(14) Note describes what the rule is for: copies matching originals. Whether a vendor's employee is a "qualified person" for a public chain the vendor does not operate, and whether 902(13) fits a public blockchain at all, are questions no authority we found resolves.
Vermont's blockchain evidence statute (12 V.S.A. §1913, as amended) creates a self-authentication path on a sworn declaration by a qualified person, and a presumption that the date recorded is the date the record was added. It then says:
"(4) A presumption does not extend to the truthfulness, validity, or legal status of the contents of the fact or record."
"(5) A person against whom the fact operates has the burden of producing evidence sufficient to support a finding that the presumed fact, record, time, or identity is not authentic as set forth on the date added to the blockchain, but the presumption does not shift to a person the burden of persuading the trier of fact that the underlying fact or record is itself accurate in what it purports to represent."
Both halves of (5) cut. The first shifts the burden of production onto whoever disputes the record. The second refuses to shift the burden of persuasion on whether the thing recorded is true.
One court has actually done it — in France, at first instance. On 20 March 2025 the Tribunal judiciaire de Marseille decided case RG 23/00046. Per EUIPO's case summary, the court gave full probative weight to blockchain timestamp reports corroborated by a bailiff's verification report, and held that those elements, together with evidence of public disclosure, established the claimant as copyright proprietor from the dates on the ledger. Limits: first instance, not binding, French procedure, fashion designs rather than music. And the evidence was anchored to Bitcoin, not Solana, the chain SongSecure uses.
Will your documentation ever reach a courtroom?
Almost certainly not. Of the 57,668 federal copyright cases terminated between 1 January 2015 and 31 March 2026, 90.1% ended with no judgment entered. Only 2.36% reached a contested judgment. Median time from filing to termination: 131 days. That is a census, not a sample: the Federal Judicial Center's Integrated Data Base. Caveats: the FJC file carries no subject-matter field, so this is copyright generally, not music specifically; and "no judgment entered" is not a settlement rate and we are not going to dress it up as one.
But the reframe it licenses is the right one: your documentation will be tested by a demand letter, a publisher's clearance check, a co-writer's differing memory, a platform takedown dispute and a settlement negotiation — all of which happen long before a judge is involved. In a negotiation, a dated record that predates the dispute and that you can produce in ten minutes does real work, not because it is unbeatable, but because the other side has to price the risk that it is true.
What each thing actually does
| Dated file records (SongSecure and similar) | US Copyright Office registration | Free open timestamping (OpenTimestamps) | |
|---|---|---|---|
| Proves a file existed by a date | Yes. SHA-256 hash on a public chain | No. It records your filing date, not your creation date | Yes. Same guarantee, Bitcoin-anchored |
| Proves you wrote it | No | No | No |
| Lets you file a US infringement suit | No | Yes, and only this does (§411(a), Fourth Estate) | No |
| Statutory damages and fees | No | Circular 1: eligible "When registration is made prior to infringement or within three months after publication" (§412) | No |
| Prima facie validity | No | Circular 1: "Registration establishes prima facie evidence of the validity of the copyright and facts stated in the certificate when registration is made before or within five years of publication." (§410(c)) | No |
| Speed | Minutes | Months. The Register must act on the application before you can sue | Minutes to hours |
| Cost at 50–100 works a year | Flat annual subscription, unlimited works | $45–$65 per single-work application, or $85 per group of up to 10 unpublished works, or $65 per album of up to 20 works; realistically $325–$850 for 100 works | Effectively free |
| Named person who will sign a sworn certification | Yes | The certificate is the Office's own record | No vendor, no affiant |
| Tested in a reported US music decision | No | Yes, constantly | No |
Note the last row. It applies to us, and to everyone selling a timestamp.
Is free timestamping as good as a paid service?
For the timestamp itself, largely yes. OpenTimestamps is free, open source and vendor-independent. It commits a hash to Bitcoin, your file never leaves your machine, and verification needs no trusted third party. If all you want is the cryptographic timestamp, that is the honest baseline, and any paid service, ours included, should have to justify itself against it. What it does not give you is file custody and retrieval, or a human who can sign a declaration under Rule 902(11) or Vermont §1913.
WIPO, the UN's own intellectual property agency, terminated its own digital timestamping service on 31 January 2022 and the tokens it issued still verify — so ask any timestamping company what happens to your proof if the company disappears.
Where SongSecure fits
SongSecure hashes your file and anchors it to a public blockchain, producing a dated, tamper-evident record of what you had and when. SendSecure, the delivery-record side of the same system, documents who received a file and when, which maps onto access, the element that killed several of the cases above.
Registration is cheaper than this category usually implies (the table above has the real numbers), so the difference is not price. What each Copyright Office option costs in 2026 is itemised in what registering a song with the Copyright Office actually costs. It is that group registration cannot cover a sequence of drafts of one song, and that the Office takes months while a hash takes minutes.
Three things it honestly is: fast, dated evidence of possession across the whole sequence; a record you can produce in a negotiation; and a named person who will sign a sworn certification, which is what FRE 902(11), 902(13)–(14) and Vermont §1913 each require, and what a bare ledger entry has no one to provide. Whether such a certification would be accepted for a public chain we do not operate has never been tested.
And what it is not. It is not a substitute for federal registration. It does not prove you wrote the song, does not make an unprotectable five-note sequence protectable, and does nothing for substantial similarity, which is what beat Bowen. It is not "court-ready," because no reported US music copyright decision has ever examined a record of this kind.
We would rather you register your songs with the Copyright Office and never use us than use us and skip registration.
What we cannot promise
- No US court has assessed a blockchain timestamp in a reported music copyright decision. Our own dataset is the evidence. Everything anyone says about how it would go is prediction.
- In the EU, the presumption attaches to a different thing. eIDAS (Reg. (EU) No 910/2014) Art. 41(2): "A qualified electronic time stamp shall enjoy the presumption of the accuracy of the date and the time it indicates and the integrity of the data to which the date and time are bound." Art. 41(1): an ordinary electronic timestamp "shall not be denied legal effect and admissibility as evidence in legal proceedings solely on the grounds that it is in an electronic form or that it does not meet the requirements of the qualified electronic time stamp." Art. 42 sets the requirements a timestamp must meet to be "qualified"; we are not aware of any public-chain anchor that has been assessed against it.
- There is a published playbook for keeping blockchain evidence out. Alexandra Sahara, Blockchain Evidence: How Smart Litigators Can Keep It Out at Trial, 92 Fordham L. Rev. Online 42 (2024), argues that forking undermines reliability. It is a student Note, so treat it as argument rather than authority. And its worked example of forking risk is Solana, the chain SongSecure anchors to.
- A ledger's immutability says nothing about the truth of what was recorded. File metadata can be forged before hashing, and the chain will faithfully record the forgery. The timestamp is trustworthy; the payload is only as trustworthy as the person who made it.
- Our dataset has known limits. 197 decisions is a near-census of written merits decisions, roughly a 2.4% sliver of what happens to copyright claims, and the pro se count was hand-identified by a single reader.
What is actually in your control
Four things, and no caveats are attached to these.
Keep the sequence. The voice memo, the session file, the lyric draft, the stems, the master — in order, undeleted. Whoever takes your song does not have the twelve things that came before it.
Write the split sheet at the session, while everyone still agrees.
Keep a delivery record of every file you send, and to whom, by name.
Register your work, early, on the form that matches the claim.
None of that needs a lawyer, a subscription, money or a blockchain. All of it is available to you this afternoon, and all of it works.
The short version
A blockchain timestamp proves you had the file. It does not prove you wrote the song.
What converts one into the other is not better cryptography. It is a sequence of dated artefacts instead of a single one, a human prepared to swear to it, and, in the US, a registration the Copyright Office has acted on, on the form that matches the claim.
We read 197 US music copyright decisions looking for a case that turned on a cryptographic proof of creation. There wasn't one. What decided those cases was what the parties could produce and date: an email whose date was three years wrong, a handwritten lead sheet date-stamped 25 October 1960, and a producer who could hand over every audio file.
Be the person who can hand over every file.
If it has already happened to you, start here instead: someone stole my song: what can I actually do about it.
Legal disclaimer
This article is informational and is not legal advice. It quotes decisions, statutes, rules and official guidance; it does not tell you how any of them apply to your situation, and no reader should act on it without consulting a qualified attorney in the relevant jurisdiction. Copyright law differs by country and changes over time. SongSecure is not a law firm. No certificate, timestamp or record, from SongSecure or anyone else, is a guarantee of any legal outcome.
Sources
Primary authority. 17 U.S.C. §§204(a), 410(c), 411(a), 411(b), 412 · Fourth Estate Public Benefit Corp. v. Wall-Street.com, LLC, 586 U.S. 296 (2019) · Unicolors, Inc. v. H&M Hennes & Mauritz, L.P., 595 U.S. 178 (2022) · Roberts v. Gordy, 877 F.3d 1024 (11th Cir. 2017) · FRE 902 and the Advisory Committee Notes to 902(13)–(14) · 12 V.S.A. §1913 (as amended) · eIDAS, Reg. (EU) No 910/2014, Art. 41–42 · US Copyright Office, Copyright Basics (Circular 1) and Musical Works, Sound Recordings & Copyright (M-309) · USCO fee schedule and group registration pages · GOV.UK, How copyright protects your work · Tribunal judiciaire de Marseille, RG 23/00046 (20 March 2025), via EUIPO.
Decisions quoted, all from the SongSecure frozen dataset (2015–2026). Bowen v. Paisley · Loomis v. Cornish · Batiste v. Lewis · We Shall Overcome Foundation v. The Richmond Organization · Nwosuocha v. Glover · Pickett v. Migos Touring · Stringer v. Richard · Buchanan v. Sony Music · Roberts v. Gordy · New Day Worldwide v. Taylor Swift · Prunty v. Vivendi · Cates v. Procter & Gamble.
Data. SongSecure frozen decision universe: 203 rows covering 197 distinct decisions (six X-DUPE, excluded), 201 retrievable as text, 108 in-scope merits decisions across 82 disputes, two independent coders, κ = 0.8297 (n = 203, 95% CI 0.7502–0.9010), measuring the inclusion screen, not any outcome variable. Federal Judicial Center Integrated Data Base — Civil: 57,668 federal copyright terminations, 1 Jan 2015 – 31 Mar 2026; census, not sample; copyright generally, not music specifically.
Commentary. A. Sahara, Blockchain Evidence, 92 Fordham L. Rev. Online 42 (2024), a student Note cited as argument not authority · OpenTimestamps · WIPO PROOF discontinuation notice · Perplexity and Bing AI answers, captured 7 August 2026.