What is already built
Not a roadmap. Everything below runs on Solana mainnet today, with real money, and most of it can be checked from the outside — on chain, on crates.io, in a verified build.
01 What it gives you
Your payroll stops being public
Salaries, contractor rates, supplier terms — every one of them is a negotiating position, and on a normal chain all of them are readable forever by anyone, including the person you are about to negotiate with.
One address, forever
You publish a single wallet and never rotate it. Payments arrive without anyone being able to total them up, and you find them by scanning the chain with your own key — nothing is handed to you, so nothing can be intercepted on the way.
Disclosure without surrender
The sender names a reader — an accountant, an auditor, a regulator — and that reader sees the date, the amount and the purpose. They can never spend the money and never freeze it. Audit without seizure.
Privacy as a dependency
Seven crates on crates.io and a reference program that pays privately in about thirty lines of Rust. You do not host a prover, run a ceremony, or understand a circuit. You add a dependency.
02 The privacy itself
03 Ways to actually use it
cargo install tidex6-mcp-local, the key stays
on your machine, limits are enforced in code, and the amounts are a fixed
set — an injected instruction cannot express “send everything”.
04 Why you can check instead of trusting
05 Live numbers
- on mainnet
- 46 payments, all received
- envelopes on chain
- 105
- disclosed to an auditor
- 18
- programs deployed
- 4 verifier, two pools, registry
- published crates
- 7 on crates.io
- ceremony
- 4 of 20+ contributors
Three independent paths were exercised end to end on the same day in August — command line, browser with a wallet, and a third-party AI agent — each completing send, audit and receive with real money.
06 How long this took
Four months, and the useful way to read them is by what was forcing the work.
No repository, no line of code. The project did not exist.
164 commits. The zero-knowledge pool worked and the first verifier went to mainnet — but payments were in SOL, amounts were public, and the sender had to hand a secret note to the recipient somehow. It proved the idea. It was not something you would ask a colleague to use.
326 commits across three repositories — twice the hackathon itself, with nothing forcing it. SOL became USDC and USDT. Public amounts became hidden ones through confidential transfers. Handing over a note became stealth scanning, so nothing is transmitted at all. The memo encryption was replaced with a post-quantum one.
Everything on this page above — agents, the relayer, in-browser proving, the reader registry, the public ceremony, the immutable verifier — was built in this stretch.
The ceremony, from 4 contributors to 20-plus, which turns the development key into a real one. A rewritten indexer so proving stays fast as the pool grows. And the first integrations by people who are not us — the only number on this page that we cannot produce ourselves.
07 What is not finished
A list of accomplishments with nothing unfinished in it is an advertisement. These are the three things we would ask about if we were reading this page.
The ceremony is not closed
Four contributors of the twenty-plus we want. Until that is done the verifying key is still a development key, and we say so here rather than in a footnote. This is the one thing that cannot be solved by writing code — it needs people, and you can be one of them.
The operator sees the send side
Wrapping into confidential tokens requires the mint authority, so the operator knows which wallet paid how much. Privacy from the public is complete; privacy from the operator on the send path is not.
There are no outside users yet
Those 46 payments are our own, made with real money to prove the thing works rather than to claim traction. Inviting strangers onto a development key would be dishonest, which is why the ceremony comes first.