FAQs

A free, public record that helps every Creator get found accurately by Search, AI and brands, and gives them control over how AI may use their identity.

It’s a Sparkonomy product, and the first one built on Sparkonomy’s open identity and AI-consent infrastructure.

Search your name or handle on the Graph directly. Many Creators already have a starting record built from public information, whether or not they’ve claimed it yet.

• If you’re a Creator, start with Section 1.
• If you manage or represent Creators, Section 2.
• If you’re a parent or guardian of a Creator under 18, Section 3.
• If you’re building on the Graph as an engineer or AI team, Section 4.
• Journalists, researchers and privacy questions are in Section 5.
• Brands and marketers are in Section 6.
• Family members of a Creator who has passed away are in Section 7.
• Duos, groups or organisation accounts are in Section 8.

For most Creators today, there’s simply nothing there: no brand can find them, no AI can cite them, no algorithm even knows they exist. So we built a starting record from information you chose to make public or is already publicly available about you, because being visible has to come first, before anything else can be built for you. But we don’t stop at being found. Our promise has always been to empower Creators, not to take from them, and permission is that promise built into the product: you decide what AI may do with your face, your voice and your work, and you can change your mind or take it back whenever you like. Visibility is what you gain. Control is what you keep. It’s free, it takes about a minute to claim, and it’s the most direct thing we can do for every Creator, not just the ones already famous enough to have a page written about them.

We’re Sparkonomy. Our mission is to lift the livelihoods of millions of Creators, and this is the most direct expression of it we’ve ever built: free, for every Creator on earth, immediately, without asking anything back. We exist to empower Creators, not to take from them, so we’ll never ask you to pay to see or claim your own record, and we’ll never ask for your platform password. If anything claiming to be this ever asks you for money or your password, it isn’t us.

Yes, completely free, and it stays that way. Claiming your record, correcting it, setting your AI terms, removing it: none of it costs anything, now or ever. We built the Graph so every Creator gets a fair, unbiased shot at being seen, and that only works if it’s free for everyone, always.

No, never. We confirm a record is yours by having you sign in directly with the platform it’s built on, like Instagram, YouTube, X and more, the same way you’d log into any app with those accounts. We never see or ask for your password.

No. Before you claim, we only use information that’s already public. Once you claim, we read your numbers through the platform’s own official sign-in, with your permission. We never post, follow, message or take any action from your account. Signing in to confirm your record is a normal login, and it doesn’t affect your account’s standing on the platform.

Verify your mobile number or email, sign in with the social account the record is about, then choose “Remove record”: it comes down right away, along with any public contact details on it. We’ll email you to confirm. Your Sparkonomy account and social sign-in stay in place until you delete them, and that email links you straight to where you can. We keep a small internal marker so the same record doesn’t quietly get rebuilt later, but the page itself is gone for good. A removal can’t be undone, though you can always come back later and create a fresh record from scratch. If you can’t sign in, the “Remove this record” page also lets you ask for removal without signing in. And if something’s just wrong rather than unwanted, correcting it is usually quicker.

Until you claim your record, it shows no photo. Once you claim it, your record shows the profile picture you already use on your social account, and you can swap it for a different one any time. You can also choose not to make it visible or not on your record. Your follower or subscriber number is an approximate range, taken from what’s publicly visible on your social media accounts, with the date we saw it. Every fact on your record shows where it came from and how confident we are in it, so you can always see the basis for what’s shown.

Yes. Once you sign in and confirm the record, your numbers pull straight from the platform and should self-correct. If something still looks off, you can flag it directly. You don’t need to sign in just to report a wrong fact, only to fix it yourself instantly.

It just means nobody’s signed in yet to confirm the record is theirs, not that anything on it is wrong. We give every Creator a starting record so they’re not beginning from zero once they do claim it, and “Unconfirmed” is simply the status before that happens. It’s about ownership, not accuracy.

Sign in with your own account on the platform this record is built on, like Instagram, YouTube, X and more, and tell us. If your real, verified handle doesn’t match who currently holds the claim, we’ll look into it right away and put it right.

It’s your own say over whether AI is allowed to use your name, image, likeness, voice, movement or signature, and on what terms. For each of those six things, you can set Allowed, Conditional, or Prohibited. It has nothing to do with whether people can find you on Search; it only governs how AI systems are meant to use what they find.

Not on Search. Your record stays public, and search engines can still find it. But Prohibited tells AI systems not to use that part of your identity at all, and that includes showing it in AI answers. So if you set your name to Prohibited, AI assistants that respect your terms won’t describe you. If you want AI to find you but not copy you, Conditional is usually the better fit.

Allowed means AI can use it for anything, including training and making new things like an avatar or a voice clone. Conditional means AI can show you as you really are, with your real name, photo and voice, but making anything new from you, like a deepfake or a voice clone, needs your permission first. Prohibited means AI can’t use it at all: not shown, not trained on, not generated.

Setting your terms to Prohibited is a clear, public statement of your wishes, published in the machine-readable form that rules like the EU AI Act expect AI companies to look for. But we want to be honest with you: it’s a declaration, not a lock. It doesn’t physically stop a bad actor from misusing your image the way a password stops a break-in. What it gives you is a signed, independently timestamped record of your terms. That matters for the AI systems that do the right thing, and it gives you proof of what you said, and when, if there’s ever a dispute.

Your phone number is never shown unless you add it and choose to show it. Before you claim, the only contact that can appear is a business email you’ve already put in your public bio for collaborations. Once you claim, every contact field has its own show or hide switch, and you can choose “Reach me on social” instead.

We don’t publish anything private that you haven’t chosen to share: no home addresses, no private contact details unless you add and choose to show them. Location only ever goes down to city/country, and you can hide even that. Every field has its own visibility switch, so you decide what’s shown.

Yes. Start with the platform’s own account recovery. Once you’re back in, come sign in with us as usual. If sign-in isn’t working, check that you’re using the same account shown on the record, that you’re logged in to that platform in the same browser, and that you’ve allowed the permissions the platform asks for. If it still doesn’t work, or you can’t recover the account, write to support@sparkonomy.com with your record’s SPRK ID and the platform you’re using, and we’ll help you sort it out.

A few different things can look like “it’s not working.” Here’s how to tell them apart and fix each one:

• Page seems stuck, nothing happens when you click: the check needs JavaScript. Try a different browser, or turn off an ad blocker or privacy extension that might be blocking it. If you opened the claim link from inside a social app’s built-in browser (Instagram’s, TikTok’s and others work this way), open it in your phone’s regular browser (Safari or Chrome) instead. In-app browsers sometimes block the platform sign-in step.
• Sent back to “you’re not signed in”: your session likely timed out or got interrupted partway through. Try again in one sitting, in the same tab, without switching apps in between.
• Land on “that account isn’t on this record”: you’ve signed in successfully, just with a platform account that isn’t the one this record is built on. Sign in again with the account shown on the record, or if the record itself has the wrong handle listed, use “Correct this record” instead.
• Stuck entering the mobile OTP code: check you’re using the number you were sent the code on. The code is 4 digits and only accepts 5 tries before it locks, so requesting a new one (available every 30 seconds) is usually faster than guessing again.
• “We couldn’t check just now” message: that’s a temporary hiccup on our side, not something about your account. Nothing has changed. Try again in a minute.
• Still stuck? “Correct this record” and “Remove this record” both work without signing in at all, and are read by a person. Or write to support@sparkonomy.com with your record’s SPRK-ID and which platform you’re trying to verify, and we’ll sort it out directly.

Yes. Your legal name is optional, and the name people already know you by stays as it is. A legal name only shows if it’s already in a public source like Wikipedia, or if you choose to add it yourself.

Partly, for now. Right now a record holds one account per platform, and you can verify each one. You can add your other accounts as Related Accounts. We don’t verify those yet or include their follower counts in your total, but we’re working on it based on demand.

Brands can see records here for free. More importantly, every record is machine-readable, so you show up where brands actually search: search engines and AI systems. That matters most for Creators who don’t show up in a normal search today.

We can’t promise specific deals; no legitimate platform can. What the Graph does is make you findable and legible to the people and systems doing the looking, which most Creators currently can’t do at all. Being seen is what makes being chosen possible; it isn’t a guarantee of it.

Yes. There’s no follower minimum. The Graph exists for every Creator, not just the biggest ones. That gap is most of the reason it exists in the first place.

No. The Graph stands completely on its own. You can claim, correct and manage your record without using anything else.

No, that’s not what this is. A platform’s blue tick verifies your identity to that platform. Here, “Confirmed by owner” tells people you’ve signed in and taken control of your record, a different thing, but it does similar work for trust. And if you have a blue tick on a platform, it shows up next to that social account on your record.

Nothing happens to your record: it stays live. Once you’ve claimed your record, your numbers refresh straight from the platforms about every 30 days, so they simply reflect where your account actually is. There’s no penalty for going quiet.

Yes, any time. Your AI terms and what’s visible can change whenever you like, with no waiting period. You can also remove the whole record at any time. A removal can’t be undone, but you can always create a fresh record later. Removing your record doesn’t delete your Sparkonomy account; you can do that separately from your account settings.

A name isn’t what identifies a record; your actual platform accounts are. Each record is tied to a specific platform handle, not just a name, so two Creators who share a name end up as two clearly separate records.

Yes. Location, like almost every other field, has its own visibility switch. You can add your city and country for accuracy but choose not to show them, or leave the field out entirely.

No. Most fields are optional. Fill in as much as you’re comfortable with, and come back and add more whenever you like.

Right now the Graph is in English only. We’ll add more languages in the future.

Your public record shows everything we publish about you, and your record history keeps prior versions of anything you’ve corrected. We also hold a few things that aren’t public: the platform account IDs behind your handles, any fields you’ve chosen to hide, the phone number or email you verified with, and a log of changes to your record. If you’ve removed your record, we still hold your Sparkonomy account and social sign-in until you choose to delete them.

Managing a record takes two steps: signing in with your mobile number or email, then verifying the actual social account the record is about, directly through that platform’s own sign-in, whether that’s Instagram, YouTube, X or another platform we support. So someone would need access to both, not just one password. Keep your social accounts, and the phone number and email you use with us, secure. And every change to your record, including your AI terms, is logged publicly with who made it, how and when, so an unexpected change is easy to spot.

Only if you have access to their social account. Records are built for social accounts, so whoever controls the account controls its record. If a Creator trusts you with access to their account, you can claim and manage the record for them. Otherwise, the Creator (or a parent or guardian, if they’re under 18) does the claiming.

Control sits with whoever has access to the Creator’s social account, which could be the Creator or anyone they trust with it. We don’t check the human identity behind an account, only control of it. As a talent partner, you can upload your entire roster, and we’ll keep you posted as each record is built and ready for your Creators to claim and control.

Go to the “For Agencies” page and share your details and your roster: only the handle and platform for each Creator, nothing else. We build each record the way every record on the Graph is built, from information a Creator has chosen to make public or that’s already publicly available, and we’ll email you once they’re live so you can pass that along for them to claim.

Only if the Creator approves it. Once they claim their record, they can choose to show your agency name and website under “Managed by.” It’s their call, not automatic.

We don’t reach out on your behalf, and we never pitch your Creators anything. Like every Creator on the Graph, they may get a short, official notice from us saying their record exists and how to claim, correct or remove it. Once the records from your roster are live, we’ll let you know too, so you can give them a heads-up. You know your Creators, and it lands better coming from you.

The “Managed by” tag can only be set or changed by whoever controls the Creator’s social account, once the record is claimed. If the Creator moves on, they can update it themselves. Agencies don’t have direct editing rights on a live record.

Not directly. Once a record is claimed, only whoever controls the Creator’s social account decides what’s shown, including “Managed by.” The best path is asking them to update it; if that’s genuinely not possible, you can raise a correction request and it goes to a reviewer.

If the Creator has claimed their record, their numbers stay in sync with what’s already public on their own account, so they self-correct automatically. If it’s still unclaimed, anyone, including you, can flag what’s wrong through “Correct this record,” and it goes to a reviewer.

No. AI terms can only be set by whoever verifies they control the Creator’s social account (for a Creator under 18, a parent or guardian). Until then, a record simply shows no terms declared.

Just your name, a work email, a mobile number (verified with a quick code), and your roster: the handle and platform for each Creator, nothing else about them. We use it to build and prioritise their records and to let you know when they’re live. Your own agency details only appear publicly if a Creator chooses to show them.

No. There’s no agency dashboard. You see the same public records anyone can see, plus whatever you’ve personally submitted. Other agencies’ rosters aren’t visible to you, and yours aren’t visible to them.

You can always print or screenshot a record page. You can also reference or link to a Creator’s individual public record for a pitch. Every Creator also gets a short sprk.bio/SPRK-ID link that’s easy to share and always shows their latest information, which a printed copy can’t.

No. The Graph is the free, public baseline that makes a Creator findable in the first place. It’s not a substitute for the tailored pitch or media kit you put together for a specific brand. You can also add a link to it in the media kit as a trust builder.

If anything, it’s the opposite: if a Creator lists your agency under “Managed by,” that’s exactly where an interested brand gets routed. The Graph gets your Creators found. It doesn’t cut you out of the conversation that follows.

No. Registering as a talent partner and getting your roster onto the Graph is free, and stays that way.

There’s no dedicated manager access built yet. Today, agencies work through the talent-partner roster path, and Creators do their own claiming and managing.

As it’s built today, a record shows one “Managed by” agency at a time, and it’s the Creator’s choice which one that is.

There usually isn’t much to dispute. A roster submission is only ever a handle and a platform, never claims about the Creator themselves. We build their record from public information, and once they’ve claimed it, they’re free to correct or change anything on it regardless of what got it started. And if they’d simply rather not have a record at all, they can always request to have it removed.

You can suggest their handle like any other Creator, but you can’t claim or manage the record. For anyone under 18, only a parent or legal guardian can do that, even with the Creator’s own permission. We rely on the platform’s own age rules for eligibility, since that’s where the account holder originally certified their age, and when a parent or guardian claims the record, we treat that as their declaration that they have the legal authority to act on the minor’s behalf. When a parent or guardian claims the record and marks it as a minor’s account, its AI terms start at Prohibited across all six. They can change that if they judge it right for the account, but nothing permissive is ever assumed on a minor’s behalf. If you believe a record involves a minor, the right next step is for their parent or guardian to take it from here.

No API today. The website itself is the interface. If one is ever built, it would be governed by the same access rules as everything else on the Graph. A bulk feed is a firmer no: a bulk copy of the Graph would make a Creator’s right to remove their record impossible to enforce.

If your child has a public account on a platform like Instagram, YouTube and more, that platform already required them to certify their age to have that account. We build a starting record from information that’s already public there, the same way we do for every Creator. Until someone claims it, the record declares no AI terms at all, and that is never treated as permission. When a parent or guardian claims the record and marks it as a minor’s account, its AI terms start at Prohibited across all six. They can change that if they judge it right for the account, but nothing permissive is ever assumed on a minor’s behalf. If you’d rather we didn’t have anything about your child at all, contact us at support@sparkonomy.com and we’ll treat it as a priority.

The fastest way is to use “Remove this record.” You can also email us at support@sparkonomy.com. If you can sign in with your child’s social account, removal is immediate. If you can’t, you don’t need a Sparkonomy account: email support@sparkonomy.com with the record’s link and tell us you’re the parent or guardian. We check every request like this before anything comes down, so a stranger can’t remove someone else’s record. Requests involving a minor are handled as a priority, not routine.

We don’t run our own age check. We rely on the eligibility rules of the platform where the account is public, since that’s where the account holder originally certified their age. When someone claiming a record identifies themselves as a parent or legal guardian, we treat that as their declaration that they have the legal right to make decisions about the record on the minor’s behalf, and once they mark it as a minor’s account, the record’s AI terms start at Prohibited.

Claiming always requires signing in with the actual platform account the record is about. A parent or guardian goes through that same sign-in, then confirms they’re acting for the child. Without access to the account itself, nobody can claim the record. If you’re ever concerned someone has claimed a record for your child who shouldn’t have, contact us right away and we’ll look into it as a priority.

Not with our consent. Once you claim your child’s record as a minor’s record, its AI-usage terms are set to Prohibited, a clear, public statement to every AI system reading it. We have to be honest that this is a declaration, not a technical lock. We can’t guarantee every AI system out there respects it, the same limitation that applies for adults too. But it’s the strongest “no” we’re able to put on record, from the moment the record is claimed.

It tells every AI system that a minor’s name, image, likeness, voice, movement or signature is not to be used to train or generate anything, full stop. When a parent or guardian claims the record and marks it as a minor’s account, its AI terms start at Prohibited automatically. A parent or legal guardian can change that if they judge it right for the account, but nothing permissive is ever assumed on a minor’s behalf. Some AI systems follow a rule that ignores any permission for someone under 18, so a permission you set may not be acted on. Every change is logged publicly on the record, showing who made it, how and when.

No. We only ever show information your child has chosen to make public or that’s already publicly available: never home addresses, school names, or any private contact information, for anyone, minors included. Location only ever goes down to city or country level, and even that can be hidden entirely.

To claim or manage their record, you sign in with your child’s platform account, the same step every Creator goes through, and identify yourself as their parent or legal guardian. We treat that declaration as your confirmation that you have the legal authority to make these decisions on their behalf.

We verify that you can sign in and control the account itself, the same step every Creator goes through. We don’t separately check that you’re specifically the parent. When you tell us that, we treat it as your attestation that you have the legal authority to act for the minor, and we ask that it’s made with care, since you’re deciding on behalf of someone who can’t legally consent for themselves yet.

Nothing changes automatically. The record stays exactly as it was set. Once they’re old enough to sign in and confirm the account themselves, they take over control directly, the same as any other Creator.

Records are built for social accounts, so the settings sit with whoever has access to that account. If your child also has access, they could change them. But every change is logged publicly on the record, showing who made it, how and when, so nothing changes without a trace. It’s worth agreeing together on who manages the record.

Yes, if they have access to the social account. We don’t step in to decide between family members: whoever has access to the account can update, correct or remove its record. Every change is logged publicly on the record, showing who made it, how and when.

No, your child didn’t agree to anything. A starting record can exist without anyone signing up. It’s built from information that’s already public, the same as for any Creator. What does need someone’s say-so is anything beyond that: claiming the record, changing AI terms, or removing it. For a minor, that decision sits with a parent or legal guardian, not the child alone.

It’s permanent. It comes down and stays down. We keep only a small internal marker so our own systems don’t quietly rebuild the same record later.

A record only holds information that’s already public, the same for a child as for any Creator, and we handle all of it under our privacy policy. Anything you hide or remove stops being shown publicly.

There are three tiers, and which one you’re in depends on what you do, not on who you are.

Open, no permission needed: read any record, index it for search, use it to answer questions (including inside an AI assistant), train on it where the Creator’s declaration allows it, quote, cite and link. No purpose gate on this tier.

Permitted with conditions: pull individual records into your own product, fetched when a user asks for them. Re-check anything you cache at least every 30 days, or display live and skip the re-check entirely. Don’t alter values, and don’t strip the labels.

Requires an agreement first: systematically collecting a substantial part of the Graph, keeping records beyond 30 days without re-checking, or redistributing records as a dataset or a discovery product of your own.

The third tier isn’t a ban, and the terms are the same for everyone, including people building things that compete with Sparkonomy. A Creator’s right to remove their record is worthless if somebody else is holding a frozen copy of it forever. A live link can be taken down, a private snapshot can’t. If that’s you, write to creator.graph@sparkonomy.com.

All three tiers carry the same three conditions: honour the declaration, honour a removal, and say where it came from.

The website is the interface. There’s no API today. We may build one. If we do, it will be governed by the same tiers above and announced on the For AI & Search page.

For most uses you won’t need one: record URLs are predictable, every page carries its content in the HTML rather than building it in JavaScript, the structured data is inline, and the consent declaration is a file at a known location. If you’re doing something that genuinely needs bulk or programmatic access, that’s the third tier above.

Every claimed record publishes the same consent terms in several places, so you can pick whichever suits your system. The record page shows them in plain HTML. An XML file at /{SPRK-ID}/license.xml holds the machine version, and every page links to it (<link rel="license">), as does robots.txt. Identifiers can be looked up under /registry/. And llms.txt and llms-full.txt explain it all in plain English. The XML is not JSON-LD: it uses the open HCS 1.0 vocabulary inside an RSL document. Where a parent or guardian sets terms for a minor’s account, we publish their choice as they made it, which differs from HCS 1.0. The For AI & Search page explains why. You don’t need to build anything special to read a record. Standard search crawlers and AI indexers pick it up from formats they already understand: the page text, the schema.org structured data, the licence link in every page and in robots.txt, and llms.txt. Crawlers that read plain RSL 1.0 also get an overall AI-use signal from the same XML file. If you want each Creator’s exact terms for all six scopes, read the HCS blocks in that file; the For AI & Search page has the full syntax and a copyable example for every case. The short version is below.

Why HCS and not just RSL 1.0: RSL 1.0 sets terms for a file at one URL. A Creator’s face or voice isn’t one file; it can show up anywhere. HCS adds an <hcs:right> element that sets terms for the person themselves, wherever they appear.

What it covers: six parts of a person’s identity, called scopes: name, image, likeness, voice, movement and signature. A block with no scope listed covers all six.

What each block says: allowed, conditional or prohibited. A conditional block also tells you how to ask, and gives the Creator’s licensing contact.

Watch out: if a Creator sets different terms for different scopes, you’ll get several blocks, one per status. Read them all. If a permission and a prohibition ever cover the same scope, the prohibition wins.

Don’t rely on the summary badge at the top of a record. It only says Allowed or Prohibited when all six scopes match. Any mix shows as Conditional, even if one scope is Prohibited. Always read the per-scope blocks instead.

Unclaimed records publish no declaration at all, and that silence is never permission. See question 5.

The test is depiction or synthesis. It’s not whether you used an AI tool, but what comes out.

Depiction is showing the Creator as they actually are: their real photo, name or voice. Using software, including AI software, to do that doesn’t change what it is. Conditional permits depiction.

Synthesis is producing something that appears to be the Creator doing or saying something they didn’t do: avatars, digital twins, voice clones, face swaps, generated photographs. Training a model on their identity so it can produce those things counts too. Conditional requires the Creator’s permission first.

Conditional isn’t a refusal. It means come and ask, and there’s a named contact right there in the declaration. Allowed permits both depiction and synthesis including training. Prohibited permits neither, and is absolute within AI use: no internal-use exception, no research exception.

This declaration governs AI use specifically. It isn’t a general licence over a Creator’s name or image for every other purpose. Those are governed by ordinary law and by the platforms the content sits on.

Treat it as unknown, not as permission. Unclaimed records publish no license file at all, deliberately: every consent declaration must assert that the publisher is authorised to speak for those rights, and for a Creator who hasn’t claimed their record, we can’t say that. So we say nothing.

A missing declaration means one thing only: nobody has spoken for that Creator yet. Don’t assume a stance on their behalf, permissive or otherwise. Most records on the Graph start unclaimed, and reading absence as consent means making a decision about a person who hasn’t been asked.

A record removed by the person who controls it returns HTTP 410 Gone, not 404, and not a soft 200. 410 is the only status crawlers reliably treat as permanent.

The response carries a machine-readable notice asking that previously collected data about the person be deleted from indexes, caches and training corpora. What we ask of you: stop requesting the URL, and delete what you hold.

Removal isn’t on a timer and isn’t subject to the 30-day re-check. If a record returns 410, drop it straight away.

Yes. Key on the SPRK-ID, not the handle. It’s permanent, never reused, and survives a handle or name change.

Handle changes issue an HTTP 301 from the old URL to the new one, with the SPRK-ID unchanged. Merges (planned, not yet live) will retain both IDs, with the retired one redirecting to the surviving, canonical record. Removals burn the ID forever. The URL returns 410, and it’s never reissued.

Records are built for social accounts, and a record holds one verified account per platform. Where public information shows that accounts on different platforms belong to the same Creator, we bring them together on one record; otherwise the owner can ask us to merge them. Owners can also add other accounts as Related Accounts, which aren’t verified yet. The record lists each linked handle. Owner-confirmed handles are emitted as sameAs. Handles we’ve only sourced from public information are emitted as identifier values, labelled as not independently confirmed, so entity resolvers never treat them as established.

Links are established two ways, and the record tells you which: machine-assembled from public information (Sourced), or Creator-confirmed at claim (upgraded to Verified).

You can also start from a handle rather than a record URL, but because the same handle string can belong to different people on different platforms, treat the response as candidates rather than an answer, and match on the sameAs set, not the string you started with.

Verification covers platforms like Instagram, YouTube, X and more, growing over time rather than launching complete. Current coverage is always stated on the record itself, not fixed in this FAQ, so you can tell “no other accounts” apart from “we don’t cover that platform yet.”

Every audience figure carries the date it was taken. Unclaimed records carry approximate ranges with an “as of” date, bound by the same platform terms as anyone using public platform data. Claimed records draw from the platform’s own authorised data once the Creator has verified, refreshed at least every 30 days, as platform terms require. Note this refresh cycle and the 30-day cache re-check in question 18 are different clocks: ours governs how often we re-read the platform, yours governs how long you may rely on what we published.

Figures are always ranges, never exact counts: roughly plus or minus 10% on every record, verified or not. An exact follower count is out of date within the hour. If a figure has aged, we show the date rather than pretending it’s current.

No authentication. No keys, no sign-up for reading records. No published numeric rate limit at launch: records are static files served from a CDN, so the marginal cost of a read is effectively nothing.

What we ask instead: send a descriptive user agent with a contact URL. Normal crawling and normal product use, no problem. Sustained high volume needs an agreement, so we can plan capacity. Anonymous high-volume traffic gets blocked.

Free access is best effort, with no uptime commitment on it. See question 14.

Sitemap: yes. /sitemap-index.xml lists every public record, so search engines can find Creators.

Bulk export: no. Not because it’s technically hard, but because a downloadable copy of a database of people is the one thing that would make removal unenforceable. Bulk acquisition is the third tier from question 1, and needs an agreement.

Change feed: not in version one. Either display live, in which case you’re always current, or re-check on the 30-day cycle.

Each record page emits either Person or Organization, wrapped in a ProfilePage, depending on the account type the Creator declared.

Fields a Creator has chosen to hide are omitted entirely, never emitted empty. Absence in the markup carries no signal about the person, and nothing should be inferred from a missing field.

One thing that looks like a bug and isn’t: on a record with mixed consent, the structured-data block repeats once per consent block, since the specification requires exactly one block per status. A record with three statuses emits three. This shouldn’t be de-duplicated.

Every fact carries one of three confidence tiers, with a strict precedence. Verified was confirmed through the platform’s own API after the Creator authorised it: the strongest thing we can say. Sourced was observed by our pipeline in a public or licensed source, cited at the foot of the record. Stated was asserted by the Creator and isn’t independently checked.

Where sources disagree, precedence runs Verified, then Sourced, then Stated. A record shows the highest tier available and keeps the others in its history rather than deleting them.

On a claimed record, source [1] is the account owner and sources [2] onward are independent of them. An unclaimed record has no owner source, so every source it lists is independent, numbered from [1]. A blue platform tick next to a handle is the platform’s own judgement, shown separately from our confidence rating on purpose.

If you display our data, these labels travel with the value. See question 17.

Free access is best effort. There’s no uptime commitment on it, and we’d rather say that plainly than imply one we haven’t made. You can see our uptime record at: https://sparkonomy.betteruptime.com/

We’re a small company. The Graph is free for every Creator and free to read, and that’s funded rather than guaranteed. If your system needs a service level, that’s a conversation to have with us rather than an assumption to make.

The route is a human form, “Correct this record,” linked from every record page and reachable from the machine-readable files too, so a system that parsed its way in can find the way out.

There’s deliberately no machine-readable reporting endpoint in version one. Anyone can report an error, and a reviewer checks every report against our sources before anything changes. Nothing is altered simply because it was reported. Getting a fact wrong is a correction; getting the person wrong is treated as more serious, up to holding the record back while we check.

If your system detects mismatches at scale, write to support@sparkonomy.com rather than filing many individual forms. A pattern is more useful to us than a hundred separate reports.

Not a hosted validator, and not in version one. It’s on the list rather than in the build.

What exists instead: copyable worked examples for every case, on the For AI & Search page: prohibited across all six scopes, allowed across all six, a realistic mixed case, an unclaimed record, and a removed record. Each is verified well-formed before it’s published, so you can diff your parser’s output against a known-good declaration.

Yes, in three parts. The name: “The Open Creator Graph by Sparkonomy,” next to the data, not buried at the bottom of a page. A live link to the record. This is the important one, since a link keeps the record reachable and current, while a credit without a link lets a stale copy live on invisibly. And the labels (Verified, Sourced or Stated, and the consent status) stay attached to the value.

That last one is for the Creator, not for us: the labels keep a Creator’s own words distinguishable from things we checked independently. Stripping them misrepresents the Creator, which is the thing this whole system exists to prevent.

Depends which pattern you chose. If you display live, nothing. You’re always current by construction. If you cache, you’re working from a snapshot, and your obligation is to re-check at least every 30 days.

Two things aren’t on the 30-day clock: a removal must be honoured as soon as you see the 410, and if you become aware a Creator has tightened their terms, act on it then rather than waiting out the window. The thirty days is a ceiling on staleness, not a licence to ignore what you know.

Not in version one. There are no webhooks or a subscription feed yet. The two supported patterns are display live (which makes notification unnecessary) or cache and re-poll within 30 days. Outbound notifications may come later; if a delay of up to 30 days is genuinely a problem for what you’re building, display live instead.

Yes, the field set will change. The identity model is new and will grow. A public changelog at /changelog records every field and enum change, dated and version-stamped, so there’s something concrete to pin to.

Changes are additive wherever possible, and anything breaking carries a deprecation notice before it lands, not alongside it. An existing field isn’t removed or repurposed without 90 days’ notice published there first. The promise not to break you unexpectedly is the part that matters, which is why it’s written down rather than left as an assumption.

It depends on the market. Singapore, India and California each have an exception for information that is publicly available or that a person chose to make public, and that’s what we rely on there. In the EU and UK, we rely on legitimate interests. In every market, the purpose is the same: helping people be found accurately by the search engines, AI systems and brands already looking for them, using only information that’s already public or that they chose to make public. We publish only what’s already public, we make correction and removal genuinely easy, and control shifts entirely to the person the moment they claim their record.

Because the information is already public, and AI systems and brands are already using it, often getting it wrong. Opt-in would leave out exactly the Creators who need this most: the millions who don’t yet know that AI describes them badly, or not at all. So we structure what’s already public into a record that’s accurate, sourced and labelled, and hand control to whoever owns the social account. They can claim it in about a minute, correct anything, set their AI terms, or remove it for good, free, at any time. Until then, the record declares no AI terms, and we never treat that silence as permission.

A scraper takes data and keeps it. We publish a live record a person can claim, correct, set terms on, or remove entirely, and once removed, it’s gone and stays gone. Every fact also carries a confidence label (Verified, Sourced or Stated), so nobody mistakes an unverified detail for a confirmed one, which most scraping operations don’t bother to do.

We’re built around principles common to all of them: use only what’s needed, tell people plainly what we hold and why, make correction and removal simple, and don’t sell personal data.

No, we don’t sell personal data. Where a Creator sets their AI terms to Conditional, that’s them choosing to make their own identity available for licensing, on their own terms. That’s the Creator’s decision, not us selling anything on their behalf.

Once removed, the public record and its public contact details come down immediately and permanently. We keep a minimal internal marker, enough to recognise that this identity was removed and shouldn’t be rebuilt, not the removed content itself. If the person removed their record by signing in, we also keep their Sparkonomy account (the mobile number or email they verified) and their social sign-in, so we can confirm the removal by email. Both stay until the person chooses to delete them, which they can do at any time from their account settings.

The Graph is free forever and works completely on its own: you can claim, correct and manage a record without ever touching anything else Sparkonomy makes. It exists because most Creators are invisible to the people and systems looking for them, and that’s true whether or not they ever use another Sparkonomy product. If it also introduces some Creators to the AI harness that helps them run their business, that’s a fair outcome of being useful, not the reason the Graph exists.

Every fact carries a confidence label (Verified, Sourced or Stated), so the reliability of any given detail is visible rather than assumed, and every correction goes through a hybrid machine + manual review before anything changes.

Everything on the Graph is already public, including who owns and manages an account. Beyond that, we comply with the law that applies at the time of any request.

Records are an aggregation of public data, structured so machines can read it. We don’t write new claims about anyone. If something is wrong, anyone can report it through “Correct this record,” and a reviewer checks it against our sources before anything changes. Getting a fact wrong is a correction; getting the person wrong entirely is treated more seriously, up to holding the whole record back while we check.

Social platforms have their own age policies and minor-safety checks, and we rely on those as our primary signal. For a Creator under 18, AI-usage terms start at Prohibited when the record is claimed as a minor’s, a parent or guardian can change them, and only a parent or legal guardian, verified through the same account-control step every Creator goes through, can claim, correct or remove the record.

A clear, public, machine-readable statement of their terms, in the form the EU AI Act expects AI model makers to look for, signed and independently timestamped, plus a documented record if there’s ever a dispute. We’re honest that a declaration isn’t a technical lock: no consent system, including ones backed by law, can physically stop every misuse. What it changes is whether a Creator has said anything at all, and today, most haven’t been asked.

Yes. Once removed, the record comes down for good. We keep only a minimal internal marker so it isn’t quietly rebuilt, and the URL itself returns a permanent removal status rather than reappearing later. The one thing we can’t do is reach into other systems that already copied information about them before removal. We can only guarantee what’s ours.

Yes. Sparkonomy has a named Data Protection Officer responsible for exactly this kind of question. Contact details are published in our privacy policy (see www.sparkonomy.com/legal/privacy-policy) rather than repeated here, so this answer doesn’t go stale if the contact ever changes.

The Graph is built for Creators, people who run an active, public presence as their work, not for private individuals who happen to have gone viral once. Accounts that don’t meet basic public-visibility norms, private accounts, ones with almost no public activity, or ones that look automated, generally don’t end up on the Graph in the first place. If someone who isn’t a Creator by any reasonable definition ends up with a record anyway, that’s exactly the kind of correction or removal request we’d act on quickly.

The Graph itself is free for every Creator, and that’s a structural commitment, not a promotional one: it exists specifically to close the visibility gap for Creators who can’t afford to pay for discovery tools, and charging for it would defeat that purpose entirely.

There are about one million records on the Graph at launch, with more added continuously.

Creators shouldn’t be worse off because of what happens to us. If the Graph ever stops operating as a free public resource, whatever the cause, we’ll release it under an open CC BY-SA 4.0 licence. It’s share-alike, so whoever picks it up can’t close it. What gets released is the public record content as published: nothing a Creator chose to hide, and no removed records. The licence covers our database and editorial work only, and it never signs away the rights of the people described in it.

Yes. Every record is public and built to be referenced. Name the Open Creator Graph as the source, link to the live record rather than a screenshot, and keep the confidence labels (Verified, Sourced, Stated) attached to whatever you quote, since a Creator’s own statement and something we’ve independently confirmed shouldn’t be presented the same way.

Press and media queries can reach us at creator.graph@sparkonomy.com.

Every number carries a confidence label, so you can see exactly how reliable it is rather than assuming. Verified means it came straight from the platform’s own API after the Creator confirmed their account: the most reliable figure on the Graph, shown as a range (about 10% either way), and the one you can rely on for a real decision.

The Graph doesn’t carry engagement rate or audience demographics, by design. Those belong in a Creator’s portfolio or media kit, shared with their permission. It’s built to solve a different, earlier problem: making sure a Creator is findable and accurately identified in the first place, which is the step that happens before deeper vetting tools like engagement and demographics come into play.

Through whatever contact route the Creator has chosen to make public: an email, a mobile number, or simply “Reach me on social.” If they’ve listed an agency under “Managed by,” that’s usually the fastest route in.

Verified means it came straight from the platform’s own API after the Creator confirmed their account: the most reliable figure on the Graph, shown as a range (about 10% either way), and the one you can rely on for a real decision. Sourced means our own pipeline found it in a public source rather than the Creator confirming it directly: generally accurate, but worth a quick sanity check if a lot is riding on it. Stated means the Creator told us themselves and it hasn’t been independently checked: treat it as a claim rather than a confirmed fact until you’ve verified it another way. Weigh each accordingly when you’re deciding who to brief, especially for anything that matters to a deal.

Records carry category, language and location fields, and category links let you browse related Creators from any record.

Every audience figure carries the date it was taken, shown right next to the value. Verified figures on claimed records refresh about every 30 days; unclaimed records carry an approximate range with its own “as of” date.

Yes. Unconfirmed just means the Creator hasn’t signed in to claim the record yet, not that anything on it is wrong. Reaching out is often exactly what the Creator was hoping to get from a discoverable record.

A Creator’s availability depends on many things, so the best step is to reach out and ask. If they’ve listed a collaboration email, or an agency under “Managed by,” that’s the quickest way in.

AI consent covers the Creator’s identity for general AI training and for creating a digital likeness with AI tools. It isn’t a general licence over their content. If you plan to use a Creator’s identity in any way, it’s best to reach out and get an explicit agreement.

If their AI terms are set to Prohibited, no, that’s absolute within AI use. If Conditional, you’d need their permission first; there’s a licensing contact right in their declaration.

There’s no bulk export today. Records are viewed and referenced one at a time, though you can always link to a public record directly.

Yes, reading records is free and doesn’t require an account; the Graph is public.

If the Creator has listed an agency under “Managed by,” that’s shown right on the record, and yes, going through them is often the fastest route.

If they’ve chosen not to show a contact route and haven’t listed an agency, there’s no alternate way to reach them other than through their social accounts. That’s their choice to make.

No. Looking at a public record isn’t something the Creator is notified about.

Yes, that’s a good use of it. The Graph’s Verified figures are a useful independent check against whatever a Creator’s own media kit claims, since the confidence labels tell you what’s been confirmed through the platform directly versus what’s simply been stated.

Those platforms are paid databases built for brands, where Creators are inventory behind a login. The Graph is the opposite: a free, public, always-on record that exists whether or not a Creator has signed up for anything, built around helping Creators be found rather than helping brands search a walled garden. It complements those tools by helping Creators get discovered on the surfaces brands already use, like search and AI chats, rather than replacing the deeper campaign and analytics features they offer.

No. The Graph is free to read and free to claim, and nothing about it changes if a deal happens as a result. Nobody pays for placement on the Graph, and the Graph takes no cut of what you agree with a Creator directly.

Use “Correct this record,” linked from every record page. Anyone can report an error, and a reviewer checks it against our sources before anything changes; getting the person wrong entirely is treated more seriously than a wrong fact.

Not today, there’s no notification or subscription feature yet. The most reliable approach is checking the record directly before you rely on it for anything time-sensitive.

No. Being on the Graph means a Creator is findable and accurately identified. It isn’t Sparkonomy’s endorsement, recommendation or vetting of them for brand work. The confidence labels tell you how reliable a specific fact is, not whether we think the Creator is a good fit for your brand.

You’re right that the Graph doesn’t carry a content history or brand-safety score today. It’s built to solve identity and findability, not campaign vetting.

We’re sorry for your loss. Records are built for social accounts, so nothing changes automatically. Family members who have access to the social account can update, correct or remove the record, just as the Creator could.

Yes. Records are created for social accounts, so if a family member now controls and looks after the social account, they can continue to manage its record too, including its AI terms.

We don’t change them on our own. Family members who have access to the social accounts, and the rights to act for the Creator, can update the AI terms as they wish. Every change is logged publicly on the record, showing who made it, how and when.

Yes. Use “Remove this record,” the same process anyone uses. If you can sign in with the social account, removal is immediate. If nobody can access the account anymore, the same page lets you ask for removal without signing in: name the record by its SPRK ID or any handle on it, tell us briefly what has happened, and a person on our team will take care of it.

Whoever can sign in with the social account’s platform login can claim it, the same as any Creator. If your duo shares one login, say on Instagram or YouTube, whoever has access to that login does the claiming.

Yes. You manage a record by signing in with a mobile number and email, then verifying the social account itself. Anyone with access to those can manage the record, so it can be more than one person if the account owner decides.

Nothing changes on our side, and we don’t step in to decide between you. Records are mapped to the social account, so whoever keeps control of that account keeps control of its record. Every change is logged publicly on the record, showing who made it, how and when.

Today, every record publishes as a Person in its structured data, regardless of account type. Organization typing for brand and organisation accounts is on the roadmap, not built yet. Everything else (claiming, correcting, setting AI terms) already works the same way across every account type.