Snapshot 2026-09-19 · Stellar mainnet

Three of the four flagship tokenized gold assets on Stellar aren't on Stellar.

We checked each issuer against its own published /.well-known/stellar.toml. Two are not issued on the chain at all; every listing that claims otherwise is a spoof. One had no on-chain identity anywhere. This site is the measurement — including the rows we could not verify.

0 of 4
targets not real
0
impersonation domains
0
bn units, spoofed claim
0
published without proof

Findings

What the chain actually says

Each row below was resolved by fetching a document the issuer publishes on a domain the issuer controls. Nothing here is inferred from a price feed, an aggregator, or a search result.

PAXG
Paxos
Not issued on Stellar at all

No PAXG entry in Paxos’ own stellar.toml. Every PAXG listing on Stellar is a spoof — one declares 85,000,003 units, another 10,002. PAXG is an Ethereum and Solana asset.

XAUT
Tether
Not issued on Stellar at all

No XAUT entry in Tether’s stellar.toml. Listings on Stellar declare supplies of 1,000,000,000 and 613,632 units under impersonated domains. XAUt is an Ethereum and TON asset.

GLDP
Gold Protocol
No on-chain identity found

No stellar.toml, no listing, no contract. We removed the entry rather than carry a placeholder that would later be cited as fact.

PAXG
impersonation
920,000,000,000 units

A stellar.toml served from blackrock.com.se — not BlackRock — advertising 920 billion units. Alongside it: mastercard.co.com, coinbase.com.co, stellar.dtcc.network. A block-explorer search hit is not an identity.

XAUM
Matrixdock
Confirmed — the one real gold issuer

Contract and issuing account both declared in matrixdock.com’s own toml. Supply is not yet measurable by our index: Soroban has no standard supply read. Graded unverifiable rather than estimated.

BENJI / USDY
Franklin Templeton · Ondo
Confirmed and measurable

$435.2M across 1,367 accounts and $461.6M across 2,748 accounts, identities read from the issuers’ own domains. The strongest attestation discipline on the chain — and not gold.

The Index

Reserve Integrity Index

3
issuers tracked
0
rows fully verified
0.00
headline, verified only
IssuerSymbolOn-chain supplyHoldersAttestation ageConfidence
Ondo US Dollar YieldUSDY461,624,094.612,748—measured_only
Franklin Templeton — US Gov Money FundBENJI435,175,948.501,367—measured_only
Matrixdock GoldXAUM———unverifiable

The headline is zero, and that is the correct number.

Only rows that are both measured and backed by an attestation with a declared cadence reach the headline. Neither measurable row declares a cadence yet, so both grade measured_only and contribute nothing. Publishing $897M of "verified" reserves on the strength of an unread attestation page is exactly the failure this gate exists to prevent.

Explorer

Every asset, and what backs it

Loading the census…

Method

Five steps, all of them re-runnable

The value of a number is not that it is large. It is that a reader can reproduce it and get the same answer.

01

Identity before numbers

An issuer enters the registry only with an identity taken from a document it publishes on a domain it controls — its own /.well-known/stellar.toml. Not a search result, not an aggregator, not a press release.

02

Measure what the chain says

Issued supply and holder count come from Horizon’s /assets endpoint. Keyless, public, re-runnable. If the record shape changes we return an error, never a silent nil — a nil renders as "unverifiable" and would disguise a parsing defect as a data condition.

03

Grade against the issuer’s own cadence

Staleness is self-referential: we compare an issuer against the frequency it declares, not a standard we invented. A monthly attester at 70 days is late. An annual attester at 70 days is not.

04

Gate publication on evidence

Three grades. verified: measured and attested. measured_only: real supply, no cadence on file. unverifiable: we could not measure it honestly. Only verified rows reach the headline.

05

Commit the snapshot

Each snapshot is canonicalised and its hash committed on-chain. A published file can be re-verified against that hash, so a figure cannot be quietly revised after the fact.

—

What this is not

Not a price feed. Not an audit. Not a reserve ratio. This measures tokens in circulation and the evidence behind the claim backing them — the part that was missing.

Verify

Don't take our word for it

Every supply figure on this site is a read of a public endpoint. Paste it into a terminal and you will get the same number, or you will find out we are wrong.

curl -s 'https://horizon.stellar.org/assets?asset_code=USDY&asset_issuer=GAJMPX5NBOG6TQFPQGRABJEEB2YE7RFRLUKJDZAZGAD5GFX4J7TADAZ6&limit=1' | jq '.balances.authorized, .accounts.authorized'

The record is nested under balances.authorized, not a top-level amount. Read the wrong field and you get a nil — which is how a parsing bug becomes a published "unverifiable".

Tokenization created liabilities. Nobody prices them.

An issuer that publishes a reserve claim takes on an obligation. This index measures whether the obligation is backed, and how fresh the evidence is. If your asset has a row here, you already know more about how we read it than most of your customers do.

Measured on Stellar mainnet. Instruments built by Actuary.Finance, verified by vna.ai.