Airdrops, governance, rewards, and grants all assume the people on the other side are distinct. Human Passport holds that assumption up, across systems that never share a database, without asking your community to hand over who they are.
Verified accounts and credentials as of 2026-06-10. On-chain figures as of 2026-07-31.
Almost every integrity problem in a community programme is one of these two. The rest of the stack answers what may be disclosed and who may act on someone else's behalf, and those matter later.
Is this a real person?
Establish that a record belongs to a real, unique human. The check holds across systems that never share a database, so one person cannot quietly become fifty applicants, fifty submitters, or fifty voters.
Human Passport →Are they who they claim to be?
Government ID, residency, age, and screening against sanctions and politically-exposed-person lists. Each check returns a verifiable result a counterparty can trust, rather than another copy of the document.
Proof of Clean Hands →A sybil attack is not vandalism. It is a rational response to a budget that pays per account.So the fix is arithmetic, not enforcement.
Figures as of the end of 2025.
The single highest-value target for automated farming. Gate the claim on proof of personhood so allocation goes to the community you actually built rather than to whoever scripted the most addresses.
Story Protocol case study →One member, one voice. A vote that can be duplicated is not a vote, and a quorum made of copies is worse than no quorum because it looks legitimate.
Optimism Citizens House →Points, quests, and referral programmes are farmed the moment they are worth farming. Personhood turns a rewards budget back into a retention budget.
ShapeShift rewards →Quadratic and matched funding are exactly as good as their sybil resistance, because the whole mechanism rewards breadth of support. Model-based detection finds the coordinated clusters signature checks miss.
Gitcoin GG23 →You are paying for signal about real usage. Without personhood you are paying for traffic, and the data you make decisions on afterwards is fiction.
Testnet case study →Moderation does not scale against automation. Verifying humanity at the door is cheaper than removing accounts one at a time, and it does not require reading what people write.
Protect a forum →A single check is a single thing to defeat. Human Passport combines credentials the person holds, behavioural scoring over on-chain activity, and clustering across the whole programme, so beating it costs more than the reward is worth.
It does this while storing no IP addresses, names, emails, or personal details. Private by default is the architecture, not a setting somebody has to remember to turn on.
Passport Embed, verification inside your own app →Stamps
Verifiable credentials a person collects by proving something about themselves, through whichever method they are willing to use. More stamps means a higher cost of forgery for anyone trying to be fifty people.
Model-based scoring
Machine learning over on-chain behaviour, which catches the coordinated clusters that individually look fine. This is the signal that finds a farm rather than an account.
Data services
Clustering and reporting on top of the first two, so you can see the shape of an attack on your programme and answer for the decisions you made about it afterwards.
A uniqueness check needs a way to say "this person already claimed" without anyone holding a list of who claimed. Human Network is a threshold network built on a verifiable oblivious pseudorandom function, and it issues the nullifiers and identity proofs that make that possible. No single party holds the secret that would let it deanonymise the set.
An airdrop, a governance round, a rewards programme, a grants pool. Bring the mechanism and the budget behind it, and we will tell you where it leaks and what verifying humanity at the door would actually change.
Holonym Foundation, a Delaware public benefit corporation