Methodology · version 1.0 · 2026
How we score, in full — and what could change it.
A rating is only worth as much as the rules behind it. So here are ours, completely. If two different reviewers on our team score the same service, this document should force them to the same number. If you disagree with a score, you should be able to point at the exact rule you think we applied wrong.
01
What counts as "no-KYC"
We use one testable question: can a person open the service and withdraw the value they put in, without ever being required to submit government ID, a selfie, proof of address, or a real-world identity? If the answer is "yes, always," it can reach level 0. If the answer is "yes — unless the operator decides otherwise," it is not no-KYC in the honest sense, and we label it KYC-on-trigger however the marketing reads.
We judge the terms and the behavior, not the homepage banner. A "100% anonymous" badge counts for nothing if clause 14.3 reserves the right to demand ID before a payout. The banner is marketing; the clause is the contract. We score the contract.
02
The KYC levels
A single 0–4 badge, decided by the definition above. Lower is more private.
- 0
Guaranteed no-KYC
Terms explicitly and unconditionally state identity is never required — including for withdrawals, disputes, or "suspicious activity". No discretionary carve-out exists.
- 1
No-KYC in practice
No ID required to register or transact, and no incident of surprise verification on record — but the terms stop short of a hard guarantee.
- 2
KYC-on-trigger
Anonymous by default, but the service reserves the right to demand ID under defined conditions (volume thresholds, "risk", regional rules). The trapdoor is documented.
- 3
Optional / tiered KYC
Basic use is possible without ID, but core functions (higher limits, fiat, faster payouts) are gated behind verification.
- 4
Mandatory KYC
Identity verification is required to use the service at all. Listed only for contrast — never presented as no-KYC.
03
The three axes
Every point below is earned by a concrete, citable attribute. The headline score is Privacy × 50% + Trust × 30% + Reliability × 20%, then hard-capped by any Reliability rule that fires. Weights are published so you can re-derive the number yourself.
Privacy 50%
What the service can demand or collect about you.
- Identity-free registration+10
- No registration needed at all+5
- Accepts Monero+5
- Onion / I2P endpoint+5
- Strict, specific no-log policy+5
- Non-custodial by design+3
- No email / phone required+3
Trust 30%
Can it technically do what it claims — and has it, for long?
- Fully open-source+7
- Independent security / no-log audit+6
- Escrow or on-chain settlement+6
- Mature: 3+ years, clean record+5
- Peer-to-peer / decentralized+4
- Responsive, documented support+3
- Transparent ownership+2
Reliability 20% + caps
The axis nobody else scores hard enough. Built from sourced incidents — it can quietly add points, or violently cap the whole score. See §04.
04
Reliability & the freeze rules
A perfectly private service that steals your money on the way out is worthless. So Reliability can override everything. These caps are absolute.
- Confirmed fund-freeze / seizure caps Overall at 3/10
- Unreimbursed loss of user funds (hack / protocol exploit) caps Overall at 3/10
- Bait-and-switch: added KYC after deposit caps Overall at 4/10
- Pattern of withdrawal "reviews" −1 to −3, by density
- Undefined "sole discretion" ID clause blocks KYC level 0–1
- Clean record across full history no penalty
Every incident behind a cap is logged on the service's page with a date and a link to the original report (Reddit, Trustpilot, Bitcointalk, on-chain evidence, court/regulator filings). One angry review is a data point, not a verdict; a corroborated pattern is what moves a score.
05
Verification tiers
Separate from the number — this says how much we've personally stood behind it.
We tested the flow ourselves and/or the operator confirmed the facts on record. Passed repeated re-checks over time with consistent behavior.
Fully audited from public policy text and evidence, but not yet operator-confirmed or hands-on tested. The default for a new listing.
Listed largely on community reports; our own deep audit is pending or evidence is thin. Treated with caution.
Credible, corroborated evidence of seized funds, bait-and-switch KYC, or dishonest terms. Score is capped and the warning leads the page.
Service has shut down or abandoned users (e.g. unpaid refunds), on sourced facts short of proven intent. Listed as a warning, never a recommendation, and never linked. Score is capped at the bottom.
06
We never stop checking
A listing is a standing commitment, not a one-time verdict. Policies change quietly; a friendly service can turn predatory after an acquisition. So every page carries a last-reviewed date, and:
- Every service is re-audited on a fixed cadence — the higher its traffic and the more money users trust it with, the shorter the interval.
- New fund-freeze reports are triaged as they arrive and can downgrade a live score the same day.
- Material policy changes trigger an immediate re-review, and the change is added to the service's incident timeline so the history is visible.
- If we can no longer stand behind a listing, we say so on the page rather than quietly deleting it.
07
Integrity & independence
- No paid listings, no paid placement, no paid scores — ever. No service can buy its way onto the site, up the rankings, or out of a warning.
- Affiliate links are disclosed and score-blind. Where a service runs an affiliate program we may use it to fund the work — but the link is added only after scoring, and the presence or size of a commission can never change a single point. Many top-rated services pay us nothing; some poorly-rated ones would pay well.
- We rate services, not people. The site returns information about services' policies — never anyone's personal data, and never a way to identify an individual.
- We don't help anyone break the law. Privacy is a legitimate reason to prefer a service. Evading justice is not the audience we serve, and nothing here is designed for it.
08
Corrections & the right of reply
If you operate a service we've reviewed, you have a standing right of reply. Show us where the facts or the clause reading are wrong and we'll correct the record and re-score — visibly, with the change dated. What you can't do is pay to change a score, because there is no price. A public correction that makes the audit more accurate is exactly the outcome we want.
Spotted something we missed — a new freeze report, a changed policy, a service that should be listed? That's what Ask the bureau is for. Real questions become public FAQs.