Priced by the corridor, not the trade. And explicitly free for the people who check the work.
Mirror is priced per corridor per year, split between the two members. The kernel, the verifier, and the conformance vectors are free forever. Supervisor and academic access is free forever. The first five corridors ratified sign in at a fixed fee, publicly named, for the lifetime of the rail. Volume, seats, basis points, and settlement are refused in code, not in policy.
The price follows the rail, in three natural shapes.
Tier 1 makes the design verifiable. Tier 2 matches the natural sales unit. Tier 3 matches the harder engineering. None of them charge on volume.
Open core
The bander source, the verifier, the conformance vectors, the threat model, the boundary register, the protocol spec, and the offline examiner pack.
- Downloadable, sha256-pinned
- No account, no licence key, no sign-up
- Same binaries a paying bank gets
- Audit-grade offline verification
Bilateral corridor
Two banks ratify a corridor together. Mirror runs the rail. Each side pays the indicative annual fee. No fee on volume. No fee on seats.
- Operator console for the bilateral pair
- Corridor-ratification ceremony support
- Conformance certification (signed)
- Technical briefings (4 / yr)
- Threat-model updates & disclosure handling
- Operator incident response
Multilateral corridor & federation
Three or more banks on one corridor, with verdicts co-signed by independent operators, multi-operator attestation, federation, and a governance committee. Premium engineering work; priced per member-year, not per verdict.
- k-of-n co-signed verdicts across independent operators (not threshold cryptography)
- Federation across operators
- Governance committee for life of corridor
- Cross-operator Merkle inclusion proofs
- VDF timestamps (when shipped)
The first five corridors ratified under Mirror sign in at a fee that is named, fixed, and does not change for the lifetime of the rail.
Names published (with consent). Names signed into the rail's registry. Names pinned on the site under /founding-corridors. The sixth corridor pays the regular schedule; the first five never do.
Free, and never chargeable, for supervisors and researchers.
The same console, the same verifier, the same worked examiner pack. No procurement. No contract. No annual renewal. It runs in your browser, or offline from the published files, with nothing to install. Regulator adoption is the moat; free is how it gets built.
The anchor against what banks currently spend.
A large bank already runs several financial-crime vendor relationships, plus internal compliance teams, plus regulatory reporting infrastructure, plus an annual external audit. The cross-bank signal — the one Mirror's bilateral corridor is built around — is currently either absent or manually assembled from indirect signals across several systems. Mirror replaces none of those systems. It sits beside them, producing an offline-verifiable signed record at audit, for one signal. The fee structure reflects that this is a new signal, not a re-pricing of an existing one. We publish no estimate of what that manual assembly costs you today — you know your own figure. Indicative for a single Mirror corridor: €240,000 / year total, split. The comparison is yours to make.
Worked figures, tied to sliders.
Drag the controls. The simulator computes a worked estimate; the indicators show what the same sliders would cost on a volume-based vendor (the wrong incentive) and what Mirror charges instead (the right one).
What does a Mirror corridor cost your bank?
Move the dials. The fee structure is one number — per corridor per year — and it doesn't change with traffic. The contrast line shows what a per-verdict fee would add, which is the wrong-incentive fee model the billing kernel is built to refuse.
More corridors, lower per-corridor fee. Same architectural refusals.
A bank with many corridors pays a portfolio schedule, never a per-verdict adjustment. The discount kicks in at the same milestones regardless of volume, flag rate, or trade value.
| Corridors per member | Per-corridor fee | Effective annual saving | Notes |
|---|---|---|---|
| 1–10 | €120,000 | — | Indicative bilateral rate. Floor for new members. |
| 11–30 | €108,000 | −10% per corridor | Portfolio schedule ratified at MSA. Same operator, same console. |
| 31+· deep schedule | ratified | −18% to −30% | Rate ratified at MSA. Multi-year commitment unlocks further discounts (3-yr: −10%, 5-yr: −18%). |
| Founding Corridors· first 5 ever | €60,000 / for life | fixed for life of participation | The first five corridors ratified are publicly named (with consent) and pay a fee that does not change for the lifetime of the rail. |
Eight refusals, enforced in code, not in policy.
The kernel's pricing layer has no input for the items below, and refuses a rate card that names one. The corridor agreement will carry the same clause. The page mirrors the rail's design logic — the price refuses to undercut the design.
Per trade, per verdict, per shipment no_per_verdict_fee
Metering traffic would charge more for submitting more trades — and coverage is the point, not a revenue line. Refused.
Basis points on trade value no_bps_fee
Mirror never sees a trade value, so it could not compute this if it wanted to. Taking a cut of value would also charge more when the bank's customer is bigger — wrong direction.
Per seat no_seat_fee
The value is in the corridor and the archive, not in the console. Seat pricing also penalises growth in compliance team size, which is the wrong direction.
Settlement or money movement no_settlement_fee
Mirror settles nothing and moves no money. There is no fee on settlement, because there is no settlement.
Filing-decision outcome no_filing_fee
The rail refuses to hold a filing decision. There is nothing to charge for, and there must not be.
Free-to-paid conversion discount no_conversion_discount
Free means free. A free tier that converts to a paid tier turns the verifier into a sales tool, which makes it unreliable, which breaks the design.
Trust score / risk score no_score_fee
Same item from the cannot list, applied to pricing. A score becomes a de-risking instrument the moment it exists.
Volume band tier (like SWIFT) no_volume_band
Volume is a side channel. The boundary register already names it. Adding pricing tiers on volume would re-introduce the leak that the design removed.
The pricing page is load-bearing. The phrase "no fee on volume" is true in the kernel, and the corridor agreement will say it too.