Every number on the result panel comes from public chain data and a scoring function that runs in your browser. This page documents both, so a result can be checked rather than taken on trust.
Shor's algorithm, run on a sufficiently large quantum computer, derives a private key from the matching public key, and it does not reverse a hash. The whole assessment therefore reduces to one question, asked identically on every chain: is the public key needed to spend these funds already visible on-chain?
An address whose key is still behind a hash is shielded today. An address whose key has been published is exposed, permanently, and no future action can un-publish it. The only thing that changes is what is left sitting behind that key.
Exposure, rather than imminent theft. Exposure is readable from the chain right now. Whether a machine capable of acting on it arrives in 2030, 2035 or later is a forecast, and credible sources disagree. Both positions are cited below.
bc1p…. The output key is itself a tweaked public key on-chain, exposed from the moment it is funded.0xef0100… but stay key-controlled. Tools that test only for code misclassify them as contracts.EVM addresses are checked across Ethereum, Base, Arbitrum, Optimism, Polygon and BNB Chain in parallel. Exposure on any one of them exposes the key everywhere, because it is the same key, and balances are summed across all of them for the value-at-stake figure.
Bitcoin balances and spend counts come from mempool.space, falling back to Blockchair, which also supplies the first-spend timestamp behind the exposure-age term. When neither reports one, exposure age is shown as unmeasured rather than guessed.
One nuance worth stating plainly: the genesis coinbase output is a P2PK script with an exposed key, but the address commonly quoted for it is a P2PKH address that has never spent. Pasting that address returns shielded, which is correct for the address and incomplete for the coins.
Small enough to read in one sitting, and it runs client-side in
/assets/js/score.js so you can inspect the exact code that produced your
number.
| Signal | Condition | Deduction |
|---|---|---|
| Public key exposure | Visible on-chain, structurally or through reuse | −40 |
| Time exposed | Over 2 years | −10 |
| 1 to 2 years | −8 | |
| 6 to 12 months | −6 | |
| 3 to 6 months | −4 | |
| 1 to 3 months | −2 | |
| Value at stake | Over $10M | −50 |
| $1M to $10M | −40 | |
| $100K to $1M | −30 | |
| $10K to $100K | −20 | |
| $1K to $10K | −10 | |
| Above zero | −5 |
The value-at-stake deduction counts in full when the key is exposed, and at 40% when it is still shielded. A large balance behind a published key is live risk. The same balance behind a key that is still hashed is latent risk that only becomes real when the address spends, and a careful cold wallet should not be graded like a compromised one.
The result is floored at zero. Grades are A+ at 95 and above, A at 85, B at 75, C at 65, D at 35, and F below that. Risk reads very low at 90 and above, low at 75, medium at 60, high at 40, and very high below that.
Three choices in this model differ from how comparable tools score. Value at stake is measured in US dollars rather than native units, so Bitcoin, Ethereum and Solana results are directly comparable. Exposure age is applied to the score rather than only displayed. Stakes are weighted by whether the key is actually exposed, which keeps the headline verdict and the grade in agreement.
| Claim | Source |
|---|---|
| 6.04M BTC, 30.2% of supply, with exposed public keys | Glassnode, 2026 |
| Roughly 4M BTC, about 25% of supply; the earlier classic estimate | Deloitte Netherlands |
| Up to 65% of Ether in accounts with recoverable public keys | Gershteyn and Alber, 2026 |
| Under 1M noisy qubits to break RSA-2048, down from 20M in 2019 | Gidney, Google Quantum AI, 2025 |
| 1.7 to 2.3M BTC exposed and effectively un-migratable | Chaincode Labs, 2025 |
| Roughly even expert odds on a capable quantum computer by 2035 | Global Risk Institute, 2025 |
| Elliptic-curve signatures planned for disallowance after 2035 | NIST IR 8547, draft, 2024 |
| The contrary view: roughly 10,200 BTC at practically exploitable near-term risk | CoinShares, 2026 |
Gidney measures RSA-2048 rather than the elliptic-curve signatures these chains use. It is cited because it moved the resource estimates, not because it breaks Bitcoin.
Check an address against everything on this page.
Run a Check