Somewhere Else

by Jim.btc · 6 November 2026

Somewhere in the long and distinguished history of human folly, there is a corner reserved for people who pay thousands for a JPEG (everyone’s word for it, even when the file is a PNG, a GIF, a film or even, god forbid, a song) and later discover that the JPEG was never theirs, never in their wallet, and never anywhere near a blockchain. It was on a server, in the care of a company, and the bill has since gone unpaid. In that arrangement, what you own is a receipt for the JPEG, with a small note attached that says “ask them”.

I should declare an interest, in the sense that someone who has spent years building a sandcastle might be interested in the tide. I’ve spent several years building on Stacks and Bitcoin, most recently Audionals and Xtrata, and some of the art we put on Stacks is quietly turning into very expensive 404s, the internet’s polite way of saying “page not found”. In October 2026 I set out to see how far the problem went. I took the 250 biggest Stacks collections by all-time volume on Gamma and checked each one twice. First I followed the link recorded by the contract. Then I looked at what the marketplace was actually showing.

Of the 241 collections that contained art, 34 had artwork that was partly or wholly unreachable through the contract. Sixteen still looked substantially intact on Gamma because Gamma had kept a copy. Think of a photograph whose negative is missing, with the only print you can find stuck to a stranger’s fridge.

The blockchain remembered who owned what. Finding the art was another matter.

How did I get here?

For about twenty years I was a music producer who ran a recording studio, and for much of that a sound engineer and tour manager, which is the job where you discover exactly how many people can be owed money at one in the morning. I found Bitcoin on a tour bus in 2011, and it ended, as these things do, with me remortgaging my house at the end of 2013 to import a crypto ATM from Canada, a proper two-way machine that would hand you cash for your coins, at a time when most of the UK’s “Bitcoin ATMs” were an iPad in a box. As financial decisions go, it was, let’s say, committed.

The machine drew a crowd of very clever people, one of whom introduced me to smart contracts, and by 2017 I was sketching the idea I actually cared about: contracts for the music business. When a band walks off stage having played its set, the stage manager ticks one final box, the contract’s terms have been fulfilled, and everyone, band and crew, is paid in the same instant. Anyone who has ever chased a festival invoice for months will understand why I thought this was the future. I still do.

Perfect contracts, imaginary JPEGs

A few years later I was introduced to Blockstack, which became Stacks. When Stacks 2.0 launched in 2021, Boom gave people an early place to mint and experiment with NFTs. We wanted to see whether we could bring artists and an audience to this new layer on Bitcoin, so I co-founded This Is Number One as a curated NFT marketplace. We wrote some of the earliest NFT contracts on Stacks, with basic royalties built in, before SIP-009 gave NFTs a common standard.

For our Genesis collection we commissioned five one-of-a-kind works. Chemical X made one solo piece and collaborated on the other four with Cara Delevingne, Fatboy Slim, Dave Stewart of Eurythmics and Orbital. As I remember it, some of them already had substantial offers to launch on Ethereum. We asked them to accept much less and help start an art scene on a network with almost no audience for their work. We told them being there at the beginning would count for something, which, I can tell you, is a difficult pitch to make to anyone with an agent. Each artist was making their NFT debut. Every piece was somebody’s number one, and that gave the project its name.

Each artist chose a charity, and the contracts were designed to split the proceeds automatically. No invoices. No chasing. No one in a back office “looking into it”. On 10 July 2021, Cara Delevingne’s Mine sold for 18,000 STX, about $21,000 at the time, in what Gamma later recognised as the first commercial NFT sale on Stacks (contemporary coverage in Cointelegraph).

It was an exciting time, watching a new network being born and making our mark on it. I felt like a boy carving his initials into a young tree. But it also had a flaw that none of us noticed, because the flaw wasn’t in the contract. The contract was immaculate about the financial side. About the JPEGs and the films it was a good deal more relaxed. The art lived somewhere else, and “somewhere else” is the sort of phrase that ought to set off alarms in anyone who has ever asked a letting agent where their deposit is and been told it’s “being held elsewhere”.

We did have a plan. Gaia let us keep files in storage we controlled and give people a way to find them. Our developer had doubts about the support it was getting, though, and we had large high-definition audio and video files to serve. He suggested Cloudinary, which was fast and built for heavy media, and with a launch bearing down on us we said yes quickly, almost as an afterthought. We were putting out real work by real artists, at the size they made it, and we weren’t about to ask anyone to shrink their art to suit the plumbing. Sometimes pioneering has a sound, and that sound is someone typing “LFG!!!” with the confidence of a person who has never once been rugged by a hosting company.

Imagine selling someone a painting and then renting them the wall, indefinitely, out of your own pocket, while the new owner admires the brushwork and never once wonders who is paying for the plaster. That is, near enough, what happened. Cloudinary is a subscription, and what nobody said out loud (or not loudly enough to be heard over the ringing in my ears, a souvenir of twenty years of soundchecks) was that somebody would have to keep paying it for as long as the art was meant to exist. For art, that is a long time. For digital art it is worse, because a file does not fade gracefully. It works perfectly right up until the day it simply stops turning up.

That somebody was Chemical X, the artist who worked closely with This Is Number One and produced many of the earliest collections on the nascent network. For years he paid to keep work that people had already bought from going dark. Not to sell more of it, not to promote anything, just to keep the links alive. From 2021 to 2026 the cost ran into the thousands of pounds. Looking back, it seems obvious that we would end up here one day, staring at an empty frame someone paid real money for, wondering whether the art is ever coming back and whether anyone left a forwarding address.

Back in 2021, in answers drafted for a press interview, he explained why he was minting on Stacks: “Opensource, decentralised apps that give us all back control of our information and other assets can only be a good thing. ‘Can’t be evil’ is the way to go.” The contracts kept that promise. The hosting, bought in good faith, wasn’t built the same way.

Ordinals, or a short education in “forever”

Then along came Ordinals, which taught me more about permanence than anything had before. You could write the file itself into Bitcoin’s transaction history. Anyone with access to that history could recover the inscription without asking the person who uploaded it. To someone who’d just watched the other arrangement being billed monthly, that was a revelation. It was also a shock. I’ve always thought of block space as sacred: every byte goes into a ledger that every node has to carry for ever, on behalf of everyone who comes after. It also struck me that Ordinals had opened a new way to attack the network. Anyone with enough money for fees could stuff blocks with whatever they liked and leave everyone else to store it. So I learned to treat the space with respect.

The catch was cost. Fees kept rising, and I got rather good at compression, in the way tour managers get good at packing for budget airlines. My first Audionals song, Truth, needed about 100 KB. I got it down to about 30 KB, then squeezed a twelve-minute song into about 3 KB. Eventually I got an eight-and-a-half-minute piece down to a text file of under 1 KB, by storing instructions for generating the music. I still needed something faster and cheaper for working musicians, a group with a famously limited budget for transaction fees.

Which sent me back to Stacks, where I’d started, by a route with more turns than my first driving test. I built Xtrata for the music work I’d been doing on Bitcoin and discovered it could do rather more. Xtrata writes files directly onto Stacks, a Bitcoin layer, in pieces its smart contracts can read. Imagine an on-chain game checking a character’s stored abilities before opening a level. With those abilities on Stacks, the game’s contract can read them and apply the rules itself. With the same data on a website, something outside the chain has to fetch it and supply it. That distinction is what makes the data useful to other programs, and it’s also what lets Forever Twins check the files it preserves.

The day the JPEGs went

Earlier in 2026 I went back to Gamma and looked up the This Is Number One tokens. Many were blank. Just a grey square where a JPEG used to be, politely suggesting I try again later.

The owners, prices and transfer history were all there, like a perfect guest register for a hotel that had quietly been demolished. On 16 October I asked Cloudinary for one of our videos. The response said the Chemical X account was disabled. The tokens were still in people’s wallets, and people could still trade them. The address for the video led to an error message.

I sent Chemical X a draft of this piece. “That was a fun read,” he replied. “I do feel disabled as it happens…” He puts the hosting bill at about £100 a month, which, if it ran the whole five years, comes to around £6,000 (my arithmetic, not his).

The same thing happened with the images we did end up trusting to Gaia: Hiro switched off its hosted Gaia hub in June 2025, so the This Is Number One images still pointing there don’t load either. The arrangements we chose in 2021 had left somebody responsible for keeping the files alive. We had made the ownership permanent and left the storage on a rolling contract.

You can see it for yourself on Gamma. Scroll the ThisIsNumberOne (V2) page and you’ll see grey box after grey box, which is minimalism, but not by choice. The Hash Ones, 91 little pixel-block pieces we deliberately made tiny, show how much depended on which account held the file. The first 31 were uploaded in September 2021 to the same Chemical X account as the films. The other 60 went onto a second Cloudinary account in early October. That second account is still serving them, and those 60 are the only pieces on the V2 page that still show up.

The Genesis collection is stranger still. Gamma shows all five pieces, which sounds reassuring until you notice that you’re looking at stills. The metadata lists each film in two places, on IPFS and on the disabled Cloudinary account, alongside its still. The IPFS copy of Mine still played when I checked it. Without following the links, I could easily have mistaken a surviving still for a surviving film, or a dead Cloudinary account for the loss of the work itself. That was when I started wondering how many other collections were standing on somebody else’s subscription.

So I counted

The 250 are ranked by all-time volume on Gamma, and each got two checks. First I would read the token URI straight off the contract, follow it to the metadata, follow that to the image, and try a few IPFS gateways if needed. Then I loaded the images Gamma itself serves, which is how most of us actually look at NFTs. I also checked who is still hosting the IPFS files, whether the owner can still change the URI, and whether anything at all lives on chain.

None of the services involved would talk to a normal script, so the entire survey ran as code pasted into a browser tab sitting on Gamma, kept awake by a little background worker because browsers throttle tabs you aren’t looking at. Inelegant, but effective. About 40 collections needed checking by eye. Nine weren’t art, which left 241.

Here’s where they stood on 13 October 2026.

The 34 with failures on the contract route are the sixteen walking dead, five failing and thirteen unreachable. Together they account for roughly 177,000 tokens and 2.5 million STX in all-time volume, around 18% of the survey’s volume. They include three of the top ten and seven of the top 25. At the other end of the table, only 27 collections were healthy or on chain. Exactly one, Stacks Invaders, could draw its own art from its own contract. Its makers, Jack Binswitch (@jackbinswitch) and Ian (@imbanksia), wrote a contract that draws each invader itself, from the block it was minted in, so there’s no picture file out there to go missing. Hats off to them. Nearly everything else is either broken already or resting on a single host: an entire house running off one extension lead, plugged into a neighbour’s garage.

On 16 October I checked the same 34 again, with Megapont Ape Club as a control. All 34 still had failures. In those three days, two collections went from one working image in my sample to none. One of them was ours, naturally. A failed check tells you where I couldn’t get the file on those dates. An artist’s master, a private backup or a copy outside the routes I tested may still turn up.

A field guide to how digital art actually dies

Not with a rug pull, mostly. With admin.

One JSON file, thirteen collections. A whole cluster from the Stacks Art days, thirteen collections, about 1.19 million STX of volume and around 21,700 tokens, point every single token at one JSON file on stacksart.com. Go there now and you get the website instead, which is not what any of them were hoping for. Not one of the thirteen is healthy. Citadels, 2,222 tokens, shows nothing at all. Zombie Wabbits is hanging on by a single image, one file that happens to still be sitting on the old site.

Same team, two set-ups. Megapont Ape Club is the biggest Stacks collection by volume, by miles, and it’s one of the healthiest: IPFS, five separate hosts. The Megapont Robots family, five contracts and about 136,000 tokens, pointed at Megapont’s own API instead. That API now returns a “Service Suspended” page, and so does the separate server behind the Megapont Space Agency spacesuits, so the Gamma tiles are blank. The Robot Factory and Space Agency sites are still online, but I couldn’t find any of the art on them. Same team, same era, opposite outcome, purely down to where the files went.

Mark McKenzie, Megapont’s co-founder, said publicly in 2023 that he’d been funding the project himself and that it had been running at a loss. He wanted a business model that didn’t lean on the community’s money. I recognised the position. Once a project is paying its way out of one person’s pocket, every monthly bill is a decision that person has to make again.

Of the 16 collections I classified as using their own servers, nine were broken. Of the 203 I classified as using IPFS, four were broken. (The thirteen Stacks Art collections sit in a category of their own, since they share one platform’s file rather than each running a server, and all thirteen are broken.) The contrast is visible in Megapont’s own work. Ape Club had copies across five hosts. The Robots depended on the project’s API. With IPFS, an address identifies the contents of a file, so someone else can serve a copy and you can check it’s the same one. An IPFS address is less ‘it’s in the shed’ and more ‘it’s the red bike with the bent mudguard’: anyone who has it can hand it over, and you can tell straight away whether it’s yours. Which is lovely, as long as somebody still has the bike.

Nine collections hardcoded Cloudflare’s IPFS gateway into their URIs, the digital equivalent of giving directions as “turn left at the Blockbuster”. When that gateway closed (August 2024), I could still reach six collections through other routes. I could find two Boomboxes cycles only through Gamma, and I couldn’t retrieve Stacks Roots through either route. A gateway gives you a way in. If it closes, there has to be both another way in and a copy to reach.

It isn’t only the 2021 stuff. Yes, 2021 is the worst year: 28% broken, against 8% for 2022 and nothing broken from 2023 or 2024. But two collections that launched in October 2025 were already unreachable in my checks, both on their own servers, both with the URI locked so nobody can point them anywhere else. A year. That’s all they got.

The bit that actually keeps me up

When neither route works, the grey tiles tell you there’s a problem. What keeps me up is the sixteen I’ve called walking dead.

With those, the contract’s URI is dead, but Gamma still shows the art. In fifteen of them every image I sampled loaded. In the last one, eight out of ten did. They’re about 1.5 million STX of volume and roughly 28,600 tokens, and they include the third and fourth biggest collections in the survey. Bitcoin Birds, the third, has since been twinned in full. If you browse them, nothing looks wrong at all. The only reason the art is still visible is that Gamma kept its own copy. It’s the NFT equivalent of a restaurant with a spotless dining room in front of an empty kitchen.

Stacks Punks V3, number four, deserves a paragraph to itself, and gets one. Every one of its 9,523 tokens points at the same single file, and that file describes exactly one picture: one shared image, with no token numbers, no traits and no per-token art in it. The contract has, in effect, never known what any individual punk looks like. Gamma does: its copy of the art sits in a folder on its own Pinata account, which is why the punks still show up on its pages. And because the contract has no freeze, its owner could repoint that link at any time.

And Gamma’s copies aren’t spread around either: 214 of the 241 collections have their marketplace images served out of one Pinata account. Blocks, 2,500 pieces, is the scariest near miss. Its contract points at a dead file, Gamma’s index points most tokens at a dead server, and the full set of images is sitting in one IPFS folder that only one pinning account holds. I could still retrieve it. It’s one bad Tuesday away from becoming very hard to find.

The good news is that 25 of the 34 broken collections can still have their URIs changed by whoever holds the keys. That gives those collections a route to repair if someone can still find the files, and whoever holds the keys can still find the keys.

A friendly look over the fence

I ran a version of the survey across Ethereum, Base, Polygon, Arbitrum and Tezos on 16 October, starting from 301 big collections, in the spirit of someone checking the neighbours’ roof after their own has leaked. 86 of the 301 point at a project’s own web server, a cloud bucket or an app host. Of 201 Ethereum-style contracts, 127 have a function that lets someone change the URI (rather like selling a house and keeping a spare key), 35 are upgradeable, and I found a recognisable freeze function in exactly one. I dug properly into the 106 that looked riskiest: 75 fragile, 14 walking dead, 14 healthy, two unclear and one unreachable. Eight projects’ domains don’t even exist any more, and for seven of those the art only survives in an indexer’s copy, held by a stranger who was never asked.

Those 106 were picked because they looked risky, so it overstates things, like judging a country’s health by who turns up at the emergency room on a Saturday night. I wouldn’t use them to estimate how often collections fail across those chains. It does show the same vulnerabilities turning up elsewhere. A token can still trade while its artwork depends on a setting someone can change or a bill someone has to pay.

Why everyone went pixel art

This is a theory, not a finding, but once it landed I couldn’t unsee it.

Why did so much of this space go pixel art? Nostalgia, sure. A strong visual language, a generation raised on 8-bit games, all of that. But I think there’s a boring reason sitting underneath it: pixel art is tiny, and tiny is cheap to keep alive. A sprite can be just a few kilobytes. A collection of them needs far less space to store and copy than the same number of high-resolution images.

Now look at proper high-resolution work, or a feature-length film. Blocks is 2,500 images at around 1.8 MB each, roughly 4.5 GB. Funky Donuts is about 7 GB. Put that on a monthly plan and the meter never stops running. Ask Chemical X.

The gap gets starker on chain. I twinned every one of the 400 Bitcoin Birds, pixel art at 3 to 8 KB each, for about 8.5 STX all in. Doing the same for Blocks is estimated at 5,600 to 12,100 STX. Per piece, that’s around 0.02 STX for a Bird against 2 to 5 STX for a Block.

Being small didn’t save the Birds. They went down with the rest of the Stacks Art cluster, and the survey caught them among the walking dead. What being small did was make them cheap enough for one person to bring back.

So I don’t think it’s a coincidence that the artists who stuck around were often the ones with the smallest files. The bill shaped the art. Anything heavy cost more to keep, so the light work stayed. The people making big, rich work hit the wall first, and I suspect plenty of them just quietly stopped. On our own V2 page, some of the little Hash Ones had a second account serving them. The films had a much bigger bill.

None of this is an argument against small. Small is wonderful when you design for it. The Xtrata Arcade is a hall of retro cabinets you can walk around, and every game keeps its own Top 10 on chain, each entry stored with a replay of the run that set it, so anyone can watch it back and check it wasn’t faked. X Chess is a full game of chess with far more going on than it lets on. Every move of every game is written to the chain and refereed by an inscribed board, so each game leaves an auditable history. Connect a wallet and you can dig back through past tournaments, move by move. Both are inscribed on Stacks, as inscriptions #3653 and #3072.

What I’m doing about it

As I said at the start, I’m not neutral here. I built Xtrata, so read the next bit with that in mind, the way you’d read a restaurant review written by the chef.

Xtrata’s current inscription contracts put files directly into Stacks itself. No pin, no bucket, no subscription with someone’s name on it. Forever Twins is how we’re using that to rescue existing collections. The first three helper contracts, for Bitcoin Pepes, LEO Cats and Miami Degens, were written by Rafshitoshi of Fak.Fun, who helped conceive the first Forever Twins contracts, and the versions we’ve launched since have evolved from those originals. Each collection gets a helper contract, and the current helpers hold a sealed record of every token’s art: fingerprint, size, file type. Once it’s sealed, nobody can change it. Me included, which I’m told is the whole point. Anyone can inscribe a twin for any token in the collection, even if they’ve never owned it. You can, in other words, rescue a stranger’s art out of sheer good manners. The helper checks the file against the sealed record, and the twin is inscribed for the wallet that holds the original. That owner is then free to swap the original for the twin and back again via the twin contract, whenever they choose to.

As of 6 November 2026, seven helper contracts are live: Rafshitoshi’s three originals for Bitcoin Pepes, LEO Cats and Miami Degens, and four newer ones for NYC Degens, Megapont Ape Club, Bitcoin Monkeys and Bitcoin Birds. The newer helpers seal a record for every token up front, 5,820 so far, each holding a fingerprint of the original artwork, so when someone comes to inscribe a twin, the contract can check that the file matches. Some of those tokens already have twins and others are still waiting. Across all seven collections, more than 2,600 tokens now have a verified on-chain preservation copy of their artwork.

Could one person twin an entire collection? Bitcoin Birds seemed a good place to find out. It ranked third by volume in the survey, its artwork was at risk, and the files were small enough that I could cover the cost myself. So I paid for and inscribed all 400, which may be the first thing in this article to come in under budget. Every Bird now has a permanent twin, and its holder can swap between the original and the twin whenever they like.

As of November 2026, when you inscribe a file with Xtrata, the whole file goes onto Stacks. That works well for the tiny compressed tracks and music instructions that first brought me to inscriptions. The masters Chemical X has handed me are a different proposition. The Dave Stewart film is 258 MB and Orbital’s is 80 MB. That’s 338 MB for two works alone, or the weight of more than forty thousand Bitcoin Birds, even if every bird were the chunky 8 KB model. The bigger the files, the more they cost to inscribe, and the more Stacks has to carry for ever on everyone’s behalf.

Block space on Stacks deserves the same respect as on Bitcoin. A film that size probably doesn’t need to be on chain. What makes it that film, owned by that person, does: the fingerprint, the metadata, the Merkle tree and the royalty rules. Those are tiny, and they’re the part a contract has to read. So these films are a good test of whether a blockchain designed for storage could keep the full masters available while Stacks keeps the record that ties each file to its token. The hard part is binding the two so tightly that neither can ever drift away from the other.

I’d already tried a version of this on Bitcoin. On 11 January 2024, I inscribed a small HTML player for Entertainment, a song I wrote and performed in 2008 as melophonic with live visuals by a fellow Stacks and Ordinals artist, DYLE. The player lived on Bitcoin and pointed to three files on Arweave: the cover image, the seven-minute track and a video of the live performance. It worked for a time in Hiro’s Ordinals viewer. I can still download the track directly from Arweave, but Ordinals.com blocks the player’s request for it. The music is still there. Whether you can press play depends on where you open the inscription. It’s a jukebox with the record still inside and the coin slot taped over.

View the inscription on Ordinals.com · Download the audio from Arweave (MP3, 10 MB)

Arweave is the strongest candidate I’ve found so far. It’s a blockchain built for long-term storage. Part of the upfront payment goes into an endowment designed to help fund storage over time, and the protocol rewards operators for keeping historical data available. A film might go years without anyone asking to see it. I want the network to have a reason to keep it through those quiet years, long after the artist has moved on, retired or taken up pottery.

A future twin could keep its link to the original token and a sealed fingerprint on Stacks, with the file itself on a storage chain such as Arweave. Anyone who retrieves the file could check that it matches. For a film, that gives us a way to preserve the full master while keeping a record on Stacks of exactly which file belongs to the twin.

I’m working through the same idea with Audionals, the on-chain music production protocol I’ve been building. An Audional is any inscribed piece of the production process: a synth, the instructions for playing it, a recording, a sample, a stem, a plugin, even a plugin’s settings. The Audional Sequencer, the on-chain workstation I’m building, pulls them together. A song’s master could be a small file of instructions on Stacks. Open it in the Sequencer and your browser fetches the vocal takes, bass line and drums from Arweave. It can fetch a synth and the notes someone played through it as two separate Audionals, then put them together as the song plays. The person who built the synth and the person who played it can each get credit. Stacks keeps the record of what went into the song and the agreed shares, so when someone pays for a play, the contract knows who gets what. You can try the paid-play half of that today: Xtrata Music pays whoever holds a song every time someone plays it.

I’ve been running the same experiment on things built purely for fun. First came Astro Blaster 3 (#3075), a little arcade game whose high scores live in a smart contract, so the leaderboard doesn’t depend on anyone keeping a server running. Then the arcade, now rebuilt so that smart contracts run the place: one lists the games in the cabinets, others look after the high scores, the posters on the walls, the big screen and the hall itself, so new games can arrive without re-inscribing anything. The hall is a 24 KB inscription that loads its parts from a fast copy, checks each one against its fingerprint and, if anything doesn’t match, reads it straight from the chain.

Then the ParallaX Casino, which is deliberately a mixture: the room and the fruit machines run on an ordinary website, because they need to be quick to load and easy to change, and only the payments go through the chain. It runs on testnet, so the money is play money and nobody is remortgaging anything. Anyone can join in, with no wallet or crypto experience needed: the casino will create a wallet for you with your device’s passkey and let you claim testnet STX to play with. All three know exactly which parts live where, and that’s the whole trick.

The test for any future option is simple. If Xtrata disappeared tomorrow, could someone still find the work, recover the file and prove it’s the same one we preserved? For the files we’ve inscribed directly on Stacks, the answer is yes. The file and the record needed to verify it are both on Stacks. Anyone can recover them from the chain without going through us. Any new storage option has to pass that same test.

The moral of the story

Nearly every failure in this article has the same root. The token sat on a chain built to outlast us all, like a cathedral built by generations. The art sat somewhere else, chosen like an afterthought. It relied on a subscription, a server, a pinning account, a gateway or sometimes just an artist’s goodwill. The chain did its job. Too often, the somewhere-else did not.

Imagine someone opens your collection five years from now. The team has moved on, the hosting account has closed, and there’s nobody left to ask. Can that collector still find the art, recover a copy and check that it’s the work they bought? If you’re launching a collection, answer that before you sell the first piece.

We’ve gone to extraordinary lengths to make sure the record of who bought the art survives. The art deserves the same care. And if your collection is already out there, make a verified copy while the server is still picking up the phone.

What to do on Monday

1. Check where the art lives. For Stacks collections, paste an NFT link, contract address or public wallet address into our free tool. It shows where the artwork and metadata are stored and flags the risks it finds. Each result is a dated snapshot. NFT Check
2. Keep the originals. Save the original files somewhere you control. (Not a folder called final_Final_FINAL_v3 on a laptop you’re about to sell.) If you’re collecting, ask the project where the token points, who keeps the files available and who can change the link.
3. Preserve a copy while you can. If you already have a collection on Stacks, you can use Forever Twins to tie a verified copy of the artwork to the original token. Forever Twins
4. On Ethereum or another chain, follow the token’s link. Look up the token URI in a block explorer. That’s the address the contract gives you for the token’s metadata. Follow it through to the artwork and see where the file actually lives. If you’d like to explore a verified copy tied to the original token, tell us which collection you care about. That will help shape what we build next. @XtrataLayers

The receipt said “ask them”. I did. Somewhere, Cloudinary will keep politely telling anyone who asks that chemical-x is disabled. For the record, he isn’t. He’s just waiting for his films to come home and, joking aside, I’d like them to outlast the error message.

Inscribed on Stacks with Xtrata. The text and all 24 images are stored on chain, each image as an inscription of its own: hover over an image to see its number. Each image also carries the fingerprint of the inscription it must match, so anyone can check the bytes against the chain.

More from Xtrata: Arcade (#3653) · X Chess (#3072) · Music · Casino · xtrata.xyz. Links to xtrata.xyz are conveniences: the inscriptions themselves, including this article and its images, can be read straight from the chain by their numbers.