World A · CEX
What CEX gives you.
Sub-millisecond matching. Deep liquidity. Cross-margin across positions. Institutional reliability. UX you can trade in.
CONVERGENCE
Two worlds.
One mark.
The performance of a CEX. The verifiability of a DEX. Inanomo's foundation — determinism, double-entry, an auditable tape — is what makes the convergence possible. Self-custody is the destination, not the starting line.
THE THESIS
Performance and
verifiability are not opposed.
The industry frame — "you pick speed or you pick trust" — is a relic of compromised foundations. With the right engine, the choice disappears.
World B · DEX
What DEX gives you.
Conservation by construction. Tamper-evident audit. Self-sovereign holdings. A trust model that doesn't require trusting the operator.
Inanomo
What Inanomo intends to give you.
The first column today. The second column on a phased roadmap. The foundation — Rust + WAL + double-entry + blake3 — is what makes the path possible.
Label: Honest phasing — see /roadmap.
TRADE-OFFS
The old trade-off, dissolved.
Speed here, trust there — that was the industry choice. Set the three worlds side by side and the compromise stops being necessary.
| Trade-off | CEX | DEX | Inanomo |
|---|---|---|---|
| Matching latency | Sub-millisecond | Block-time | In-process path |
| Custody of funds | Operator-held | Self-custodied | Threshold now, self-custody phased |
| Verifiability | Trust between audits | On-chain | Audit chain + off-box tip |
| Listing gates | Curated | Permissionless | Curated at launch, phased |
The last column is the honest one: the CEX-grade engine is Live today, while self-custody and on-chain settlement are phased and clearly labelled on the roadmap — not claimed as shipped.
WHERE IT LIVES
Each part sits where it belongs.
Not everything wants to be on-chain, and not everything should be off it. Three pieces, three homes — each with a way to check it.
01 · Off-chain matching
Deterministic matching, off the chain.
The matching core runs in-process on a single-threaded Rust actor — deterministic and fast, with no block cadence to wait on. State is a bit-identical replay of the write-ahead log, not a recomputation.
Check it: numbers are published with their methodology — in-process matching path, single instrument, canary hardware.
02 · Audit chain
A tamper-evident tape, exported off-box.
Every fill, transfer and balance change is hashed into an append-only blake3 chain. The latest hash — the tip — is exported off the machine that produces it, so history cannot be quietly rewritten.
Check it: the off-box hash-tip is a commitment someone else already holds.
03 · On-chain anchor
A public anchor, only when you want one.
For a collection that wants a public checkpoint, a later option anchors its Merkle root on-chain — one published commitment to the off-chain history, without paying gas on every trade. It is opt-in.
Check it: the verifiable ledger stands on its own — the anchor only adds a blockchain hook. Planned.
DETAILS
Questions, answered straight.
Whether this is a DEX, where funds live, what is verifiable today, and why not everything is on-chain.
Is Inanomo a DEX?
Not today, and the page does not pretend otherwise. What runs now is a CEX-grade engine — a deterministic Rust matching core with a double-entry ledger and an auditable tape. It borrows the DEX’s verifiability through that tamper-evident chain, and self-custody is the phased destination, clearly labelled on the roadmap rather than claimed as shipped.
Where do my funds live?
On one canonical double-entry ledger, expressed in three roles — Custody, Trading and Margin — with funds at rest held under threshold-signature custody, where no single machine can sign a withdrawal alone. Self-custody is the destination the roadmap works toward, not the starting line.
What can be verified today, and what is planned?
Today the marketplace asset-ledger can be replayed end to end by an external verifier, and on the exchange side the blake3 audit chain and the conservation gate run continuously with the hash-tip exported off-box so it cannot be silently rewritten. Full public proof surfaces that let you re-run those checks yourself are Planned for Phase 02 — that boundary is stated plainly.
Why not just build it fully on-chain?
Because matching on a chain inherits the chain’s block cadence, and that is a latency a serious book cannot afford. Running the matching core off-chain keeps it deterministic and fast, while the audit chain and conservation gate carry the verifiability — and for anyone who still wants a public reference, an on-chain Merkle-root anchor is an opt-in option, not a per-trade tax.
KEEP READING
Where the convergence is built.
The engine that makes it fast, the receipts that make it verifiable, and a ledger you can already replay.