Planning baseline: 5bdb9b03c5e5b3551c4735d6d39dad2ccddd0a4b, inspected on
2026-09-04. This specification preserves the launch objective; it is not release
approval, current test evidence, a customer commitment, or approval to spend funds.
The wedge is small USDT treasury desks repeatedly moving working capital between Tron payouts and Ethereum DeFi. Start with a complete public-testnet pilot, then admit real funds only after the applicable release and risk controls pass. The product is instant private payments and swaps, including cross-jurisdiction swaps. Payment-only is not a launch milestone.
The smallest credible deployment has three operationally independent hub owners, H1/H2/H3, and market-maker liquidity. Each hub owner runs one Runtime containing its Ethereum and Tron sibling Entities. Each user and MM similarly keeps its own two sibling Entities together in its own Runtime. Operators have separate keys, hosts, administration and recovery ownership. Three children controlled by one operator do not demonstrate independence.
Offer Ethereum USDT, Tron USDT and one liquid Ethereum pair beyond USDT, selected from the first desks’ actual demand and executable MM quotes. Asset identity is jurisdiction-local; identical ticker names never make the two USDT assets interchangeable. Use Ethereum Sepolia and Tron Nile for the public-testnet loop; Ethereum mainnet and Tron mainnet remain the intended real-funds rails.
The first user must complete funding, a private payment, a same-jurisdiction swap, a cross-jurisdiction swap, and withdrawal through the app. Include cancellation, restart and recovery drills. Funded off-chain completion is measured separately from chain funding and finality. Publish the privacy boundary accurately: private bilateral state and encrypted transport do not justify an unconditional anonymity claim.
The landing page presents that workflow, current network stage, supported assets, fees, liquidity and recovery coverage. Foundation endorsement is the default discovery view, with a visible permissionless view for unendorsed hubs and jurisdictions. Signature verification, endorsement and operational health have distinct meanings.
The sharpest adoption risk is asking desks to prefund unfamiliar hubs before dependable quotes, liquidity and exits are demonstrated. The pilot must earn that working-capital allocation through repeated successful use.
Interview 15 desks about an existing recurring Tron/Ethereum treasury workflow; recruit five design partners with a named operator and a real recurring task. Approach two OTC or payout businesses as referral partners and two MMs as liquidity partners. These are recruitment targets, not claimed relationships. Seek switching evidence in lower total cost, faster usable funds and fewer manual steps.
Onboard each desk with a guided testnet round trip, an agreed trade-size band, a quote comparison against its existing workflow, and a recovery/exit rehearsal. After real-funds approval, start with that same bounded workflow. Do not count a faucet-funded rehearsal as a funded customer or business volume.
Bootstrap liquidity with explicit MM inventory commitments on both jurisdictions, two-sided quotes at agreed sizes, disclosed spreads, and measured replenishment times. Keep MM inventory distinct from operating capital. Incentives reward completed, attributable customer activity and sustained usable quotes; exclude self-trades and manufactured round trips. Do not promise uncapped inventory or liquidity guarantees.
Owner-approved pricing (2026-09-06): zero platform fees for the first 30 days, then 1 basis point per completed payment and 3 basis points per completed swap, plus disclosed hub fees, spread and chain gas. Show the total quote before acceptance. The business terms are approved; canonical fee accounting, the configured fee recipient and actual deployment still need implementation evidence. Testnet can show the approved fee schedule without booking fixture activity as real revenue.
The useful intermediaries are the two distribution partners, independent hub operators, MMs, recovery/watchtower providers, and reviewers who can verify the deployed contract/runtime evidence. Foundation endorsement helps discovery but is not a guarantee against loss. Before real-funds rollout, establish the applicable legal and operational responsibilities for the selected desks and operators with qualified advisers. Do not start by pursuing broad bank or CBDC partnerships.
The commercial clock starts when the real-funds pilot is approved and usable. Public-testnet recruitment and rehearsal results are reported separately.
| Milestone | Commercial target | Evidence |
|---|---|---|
| Day 30 | Five funded desks; 100 genuine completed operations; quotes available for at least 95% of requests at agreed sizes | Completed end-user operations, both payment and swap use, and all quote requests including unavailable quotes |
| Day 60 | Ten weekly active desks; at least 50% four-week retention; at least 30% lower measured all-in cost | Cohort history and like-for-like comparisons including spread, hub fees, platform fees and gas |
| Day 90 | Twenty weekly active desks; two partners each originating three active desks; positive contribution after subsidies | Attributed active desks and revenue less attributable variable costs and incentives |
A weekly active desk completes a genuine business payment or swap in that week; report payment users, swap users and users of both separately. Four-week retention uses the original activated cohort, without replacing churned desks in its denominator. Compare cost at the same trade size and destination asset, and show funding/rebalancing costs separately from the funded off-chain path.
Continue only with zero unexplained loss, at least 99% completion of accepted quotes, and funded off-chain completion p95 at or below two seconds. Quote availability has all requests in its denominator; completion has accepted quotes in its denominator. Report timeouts and user cancellations explicitly.
Pause money movement on an accounting or exit fault, preserve evidence and repair the first fault before resuming. Stop expansion if day-90 cohort retention remains below 50% or contribution after subsidies is non-positive. A smaller reliable pilot can continue while the economics are repaired; growth does not excuse an accounting failure.
Before a round, produce repeat retained usage, paid pricing evidence, positive contribution after subsidies, dependable liquidity at useful sizes, three independent operators, and reproducible settlement/recovery evidence. Separate customer volume from testnet, MM maintenance, subsidies and benchmark traffic. Do not fundraise on submitted transactions or an unverified TPS projection.
Likely strategic investor and customer classes are stablecoin payout/treasury providers, OTC businesses, MMs and wallet or DeFi distribution businesses that benefit from this exact flow. Approach them with the measured pilot and a concrete integration request. This is an outreach hypothesis, not a commitment from any named organization.
Use operating capital for protocol reliability and security, three-operator operations, recovery readiness and the two partner integrations. Budget liquidity inventory separately, with explicit ownership and exposure limits. No fundraising amount or equity/governance change is specified here.
Expand in this order: deepen dependable liquidity in the initial corridor; embed the workflow in the two partner products; add assets demanded by retained desks; then add qualified jurisdictions with demonstrated operations and demand. Retain Ethereum for DeFi and Tron for USDT rails throughout the initial launch.
Root implementation owns the production boundary. This is an acceptance specification, not another live TODO list; todo.md remains the release status. fints.md, the canonical architecture and AGENTS.md retain their normative authority.
| Priority | Deliverable | Acceptance evidence |
|---|---|---|
| P0 | First real Account opening and payment over production P2P with the selected TS/native Rust processes | Commit roots agree; no retired embedded authority or shadow path is restored; preserve the first failure capsule and add its smallest regression vector |
| P0 | One immutable mixed production WAL | Replay the same checkpoint/WAL through TS W1/W4 and Rust W1/W4; every R/E/A root and ordered event/effect/outbox digest agrees per frame |
| P0 | Live jurisdiction integration | Native Rust watcher → Entity → batch → authenticated receipt; strict rebalance, same-j swap and cross-j/MM flow; fix the first divergent production boundary before broadening tests |
| P1 | Complete public-testnet product | Current-byte Sepolia and Nile contracts/adapters; app funding → payment → same-j swap → cross-j swap → withdrawal; cancellation, partial fill and dispute/recovery evidence; preserve receipts and ordered roots |
| P1 | Three independent hubs and MM liquidity | Three separately operated owner Runtimes, each with two local sibling Entities; independent MM Runtime; quote depth/availability and recovery after hub/MM restart demonstrated |
| P1 | Honest discovery and onboarding | Foundation endorsement shown by default; unendorsed permissionless view remains available; signature verification does not masquerade as endorsement; explicit chain, asset, counterparty, fee and recovery presentation |
| P1 | Landing and complete browser journey | Pilot offer replaces unsupported scale/anonymity claims; real app flow completes with zero uncaught browser errors, truthful stage labels and usable recovery instructions |
| P2 | Release candidate and real-funds authorization | Current immutable candidate passes the applicable scenario, Rust production/test, browser, recovery, deployment and security gates, then bun run check; risk-control and contract blockers are resolved before claiming real-funds readiness |
| P2 | Production performance evidence | After exactness and live J gates, run the required locked live H1 baseline; count committed unique operations, drain all Account ACKs and prove zero transport loss |
Use the stand lock for every heavy run. Production TPS requires at least 1,000
active sovereign user Runtimes packed 200 per process, a full 20-second window at
at least 1,000 offered payments/second, at least 1,000 committed economic
operations, real WAL/fsync, an explicit TS/Rust engine and a five-second drain.
Replay, enqueue counts and AccountTx throughput are not TPS. Push only validated
milestones after the relevant L1/L2 evidence and bun run check.
riskCapUsd: null and
riskCapEnforcement: not_implemented. The older USD 10,000 prose is an
unenforced proposed ceiling, not protection. Independent hub funding limits do
not by themselves prove an aggregate ceiling. A real enforcement design is a
protocol/business decision if existing canonical rules do not determine it.Tron entry is a
second EVM test chain. MM mesh runs dual
Anvil; the cross-j scenario covers two
Runtimes. Neither proves live Tron or independent operators.operator-config and verified-gossip-profile role evidence, not Foundation
endorsement. Cross-j topology
requires each owner’s sibling Entities together; do not split one hub owner’s
cross-j legs across independent owner Runtimes.MML, clarified by the owner on 2026-09-30, is to make the accounts supporting 51% of world GDP provable by 2050, with dispute on underlying programmable J-machines. It measures coverage, including activity that never needs individual J settlement. Payment turnover divided by GDP does not establish that coverage; the attribution method remains to be specified.
The following operation measurements provide near-term adoption and reconciliation evidence. They are separate from GDP coverage, technical throughput and assets deposited.
Use existing payment lockId and swap/order identities, with their owner/domain
context, to reconcile end-user economic operations across Runtime records. Derive
analytics from committed evidence; do not add durable consensus fields solely
for reporting. An accepted quote, submitted input, proposal or unresolved Account
ACK is not completion.
Count a payment once when the recipient’s economic result is committed. Count a swap’s executed value once at a disclosed valuation convention, not once for each asset leg; partial fills contribute only their newly completed value. Cross-j completion requires both legs’ committed outcome at the same proven fill ratio. Retries, route hops, sibling mirrors, duplicate ACKs and later on-chain netting of already counted activity contribute zero additional value.
Exclude testnet/faucet activity, operator maintenance and liquidity rebalancing, self-owned transfers and manufactured incentive volume from commercial activity. Real customer swaps against an MM remain customer activity. Publish the valuation source and time, unique operation count, missing-evidence exclusions and reconciliation totals. Track user payments, same-j swaps, cross-j swaps, quote availability, accepted-quote completion, latency, cost, retention and contribution alongside provability coverage so gross volume cannot hide an unusable product.