The hard questions, answered straight.

Minting is free during open testing by design. The paid tiers ($6K self-serve, $25K named institutional) begin when open testing closes; The ZK engine, the self-serve portal quants submit through, browser-side seal-verification, and the marketplace engine (allocations, capital-routing, fee accounting) are all built. What's missing is what the raise funds: distribution, allocators onboarded, our data feeds, and live billing — we are pre-revenue with zero paying customers and no live payment processing. Revenue is a flat $6K/yr credential ($25K institutional tier), $25K/yr allocator seats, plus a marketplace success fee — on marketplace-routed capital the manager keeps 15%, 5% goes to the standard layer (MIZAN 2.5%, the partner venue 2.5%), and the allocator keeps 80%, crystallized annually above a high-water mark, so it follows credentialed capital rather than leading it. Free forever, by design: running the verifier and registering trials in the honest-N ledger cost nothing — charging for either would corrupt the statistics the standard exists to protect. First credential free for the founding cohort (25 managers). We'd rather show you a working product and an honest $0 than a revenue number propped up by a product that isn't real.

It is, and we won't pretend otherwise. The mitigant is the evidence: a working ZK engine, 18,000 lines of Rust, a live product, and nine fully published self-retractions (40+ retracted overall) were all shipped solo. That's the strongest possible signal on the one thing investors most want to de-risk — can this person execute. The raise funds the first hires (engineering + distribution). Dr. Anish Mohammed advises on the cryptography.

Live / built and tested, today: the ZK verification engine, the interactive studio (verify a real credential in your browser — now including the STARK seal itself, client-side, with no trusted setup), the self-serve product (signup → submission → issued credential → public per-credential page), chained tracks (a credential re-proven over a strictly longer window as data accrues — the live chain is public and re-verifiable), the marketplace engine (allocations, capital-routing, fee accounting), new asset-family engines (point-in-time equities and ML / black-box paths — verifiable AI performance, live: an AI strategy's record proven without revealing the model), a remediated internal security audit, the public retraction ledger, a registry of 77 credentials — passes and refusals — and 24/7 autonomous research. Forward (clearly labeled as such on the site): real capital flowing through the marketplace, allocators onboarded at scale, live payment processing, our licensed data feeds, and an independent third-party audit. We mark the line between built and coming on every page, never blurred.

Yes — that's chained tracks, live since 2026-07-16. The same sealed strategy is re-proven over a strictly longer data window as data accrues; each link is cryptographically bound to its parent (same strategy commitment, same locked gate, the old window a Merkle prefix of the new). Your backtest becomes a growing, verifiable track — one click in the app. The live chain: a 19,670-bar parent, committed and anchored in Bitcoin block 962,013 before the new data existed, extended to 19,671 bars on 2026-08-11 — 20 checks green across the chain walk (the 17 single-credential checks plus chain-continuity checks). Re-verify it yourself. An earlier v3-era chain (parent 800 bars → extension 1,096 bars) remains public as the first engineering proof of the mechanism. Boundary: chaining extends the proven computation window; it does not attest live execution or fills.

Three ways. (1) They require a human to trust; MIZAN requires math — anyone re-verifies a credential in seconds with no trust in the issuer. (2) They need to see your strategy; MIZAN proves performance is real while the strategy never leaves the proof (zero-knowledge). (3) They take weeks and cost five-to-six figures; MIZAN is instant and standardized. It's not a faster auditor — it's a different primitive. Think of the lineage this way: GIPS verifies the process; MIZAN proves the math.

No, and it never will — they answer different questions. GIPS governs how a firm reports the performance of real client money: every qualifying account included in its composite, prescribed disclosures, fee treatment, error-correction policy, reviewed by a human verifier. It is an organizational and presentational standard. MIZAN proves the computational integrity of a strategy: positions on committed data, costs charged inside the proof, out-of-sample, deflated against a committed trial count — verdict re-derivable by anyone, strategy never revealed. A manager can be fully GIPS-compliant and still show you a fitted backtest; a MIZAN credential says nothing about their composite discipline. GIPS verifies the process; MIZAN proves the math. Ask for both.

Two reasons. First, you never have to take our word — verify it yourself with the source-available verifier; the math doesn't care who issued it. Second, a verifier is only credible if it's willing to say FAIL: and we've publicly killed our own strategies — nine published in full, 40+ retracted overall: naming the exact bug (survivorship, lookahead, phantom fills) that fooled each backtest (the ledger). A firm that publishes its own failures is a firm whose passing claims mean something.

Yes, and at the honest bar, our own flagship fails. Every credential is scored on the Deflated Sharpe Ratio (the Bailey–López de Prado significance test that discounts a Sharpe for the number of backtests it took to find). At the 95% significance bar, MIZAN's flagship strategy reads "not significant at 95%": and we mint that verdict onto the credential, publicly, rather than bury it. It's live, GPU-proven, and Bitcoin-anchored; re-verify it yourself. A verifier that flatters its own owner is worthless. Ours doesn't, which is exactly why a PASS from it is worth having.

Those failure modes are exactly what the credential is built to catch — and the literature confirms their prevalence: models reaching 100% accuracy in-sample score roughly 50% out of sample (sources: mizan.market/evidence). Today the engine enforces point-in-time data, real after-cost returns, a forward-committed (un-backfillable) window, and the full López de Prado significance suite. As of 2026-07-21 all three of his statistical corrections are live and re-verifiable, each minted as a real STARK: the Deflated Sharpe test (selection bias — penalizes a Sharpe for the number of trials behind it), PBO (probability of backtest overfitting, held to ≤0.5), and CPCV (purged, embargoed combinatorial cross-validation for leakage). We found the bugs we retracted in our own work — that discipline is the product, not a liability. The retraction ledger is the receipt that we hunt them rather than hide them. Two honest limits still hold and we say so: the trial count N is forward-only (it can't recover experiments run before the ledger opened), and high-frequency strategies are refused outright — bar data cannot honestly verify them. Serial correlation itself is handled: the engine deflates by an AR(1)-corrected effective sample size, and since v11 the Harvey–Liu/Hansen suite (PSR, SPA, minimum backtest length) runs alongside the López de Prado corrections.

No to both. The strategy is concealed by the zero-knowledge proof — MIZAN verifies it without seeing it. There are two ways to prove, both yielding identical credentials: run the prover on your own hardware (MIZAN sees nothing, trust assumption: none), or — architecture built, live deployment pending — on MIZAN's sealed confidential-GPU enclave, where the witness travels end-to-end encrypted (HPKE) and a hardware attestation proves even MIZAN couldn't see it. Not "trust us not to peek" — proof that peeking was impossible. The attestation core is shipped and tested in the open codebase. And allocation is non-custodial: capital stays in the manager's or allocator's own brokerage account or on-chain vault. MIZAN is the verification standard, never the counterparty.

The in-circuit gate is a continuous, cryptographic audit of every credential. A formal third-party ZK security audit of the circuit is on the roadmap and funded by this round; Dr. Anish Mohammed (co-founder & CTO of Panther Protocol, Ethereum Orange Paper reviewer) advises on the cryptography. We don't claim an audit we don't have — we tell you exactly where it sits. In the meantime we adversarially reviewed our own engine and published the full findings, including a critical vulnerability we found, fixed, and regression-tested, and the items still open.

Fundraising details are shared directly with investors — [email protected]. The product is built and tested; the round closes the distance from built to live.

Thesis funds in AI, ZK, and deep-tech infrastructure who understand that verification is the durable franchise in finance — the Moody's / MSCI / ICE pattern, and that systematic finance is the last large market without one. The moat compounds: every credential is Bitcoin-anchored and forward-committed, so a copycat can't backfill history — they start at N=0 the day they launch, and stay there.