Find the cause. Prove the fix.
Deterministic audit tools and an AI review panel go through your EVM-compatible contracts or your DAML source and turn raw tool output into a findings list your team can act on.
Audit a Solidity contract
Upload a source file — deterministic tools and the AI review panel run together, correlated into one report.
Advanced: branch/tag and subfolder
Audit findings
Run an audit to turn tool output into an actionable review list.
These findings are automated evidence for a qualified engineer to review — not a certification that the contract is 100% issue-free. Confirm exploitability and remediation yourself before relying on this in production.
Load a previous report
Upload your updated contract(s) and the .json report from a previous audit. Loading is free — no run, no payment, no AI call — but every re-check round after it is paid. Just finished an audit? Use Fix & re-check on the Audit code tab instead and skip this step.
Fix, justify, re-check
Load a report on the left and this becomes the engagement: every finding, its status, and every round you have run.
Mark each finding fixed — describing what you changed, if you want it in the report — or accept the risk with a written justification. Submit updated source for a re-check round and every claim is verified rather than taken on your word, so nothing is marked resolved without being checked.
Tell us what is working, and what is not
Read by a person, not a queue. If something in a report was wrong, misleading, or missing, that is the most useful thing you can send us.
Request a model
The review panel runs several models against your contract. Tell us which one you want added and we will count the requests — the models people actually ask for are the ones worth the integration work.
Requests for the same model are grouped however you spell it, and asking again updates your note rather than adding a second vote.
A review panel of your own
Several model families reviewing in parallel, a jury that argues findings before any reach you, and rulebooks written around your protocol’s own invariants — the architecture behind our EVM and DAML reviews, built for your codebase and your definition of a finding.
Everything on this account
Audits and the points they cost, in one order. A round that failed shows what came back.
Reports are not stored — regenerating one means re-uploading the same contracts. We keep the record of the run, not your source code.
Money and activity
Everything below is derived from the points ledger, which is append-only — no figure here is stored, so all of it reconciles back to individual rows.
—
"Paid in" is cash received. "Spent on audits" is what users actually consumed — that is the revenue you have earned. "Unspent credit" is points bought and not yet used: it is a liability, not revenue, because those audits are still owed.
RECEIVING WALLET (LIVE, ON-CHAIN)
—
Read straight from the chain, and deliberately not reconciled against the ledger above: withdrawals leave the wallet without touching the ledger, and a transfer sent directly to the address arrives without any app-side record.
AUDITS BY TYPE
AUDITS BY HOW THEY WERE PAID
On-chain rounds appear here and nowhere else: their payment records are swept shortly after they settle, so this history is the only durable trace of them.
Payments on the wrong chain
Money that reached the receiving address on a network the payment was not created for. It is not credited automatically — each of these is a customer waiting.
Users
Points held, lifetime paid and spent, and audits run. Search by email, or leave blank for the newest accounts.
Discounts
Set a site-wide discount, or give one account its own rate. When both apply, the account is charged whichever is larger.
A discount reduces the audit and deep-review run price. It does not apply to the per-line size surcharge, to remediation rechecks, or to buying points — those always bill at list rate.
SITE-WIDE DISCOUNT
No site-wide discount is configured.
FREE CREDIT — ALL ACCOUNTS
Adds points to every account that exists right now. Accounts created later are not funded — this writes credit once, it is not a standing rule. Granted points spend like bought ones but are never counted as revenue.
Nothing granted in this session.
PER-ACCOUNT DISCOUNT
Set these from the Users panel above — find the account, type a percentage on its row, and apply. Every account that has one is listed below.
Contracts per audit
How many source files one audit may carry. Set a site-wide cap, or give one account its own — an account's own cap wins outright, in both directions.
The cap bounds uploads, the GitHub fetch and what a payment can be created for, together. Raising it raises the largest audit an account can be charged for; every extra file multiplies the tool suite's container runs and the AI hunt's token spend.
SITE-WIDE CAP
Loading…
PER-ACCOUNT CAPS
Set these from the Users panel above — find the account, type a number of contracts on its row, and apply. Every account with its own cap is listed below.
Minimum margin
The margin a round is expected to clear. This is a floor for reporting, not a price — nothing here changes what anyone is quoted or charged.
A round is quoted and paid before the expensive work runs, and the jury — most of what a round costs — scales with how many candidate findings a contract produces, which is unknowable at quote time. So this reports rounds that already sold below the floor rather than preventing them. To stop them happening, lower the per-round jury cap instead; the two settings are the same decision from either end.
SITE-WIDE FLOOR
Loading…
Entered as a markup: 50 means a round should sell for at least cost × 1.50. That is a 33% gross margin, not 50% — both figures are shown above so the two are never confused.
PER-ACCOUNT FLOORS
An account's own floor wins outright over the site-wide one, in both directions — so a strategic account can be served at a thinner margin deliberately. Set these from the Users panel above: find the account, type a markup percentage on its row, and apply. Every account with its own floor is listed below.
ROUNDS BELOW THE FLOOR
Measured rounds only. Rounds run before cost metering existed have no cost to compare against and are never judged.
Where the money goes
Measured cost of producing audits, split by the agent role and the model that spent it.
This is the evidence for whether the price formula is pointed at the right thing. The size surcharge assumes a longer contract costs more; if the jury roles dominate this list while cost barely moves with lines of code, then cost tracks how many findings a contract has, not how long it is — and the surcharge is indexed on the wrong variable.
Loading…
BY AGENT ROLE
BY MODEL
Configured discounts
Every discount currently on record, enabled or not.
Change history
Who changed which discount, when, and to what.
What users told us
Callback requests waiting for an answer, model requests ranked by how many people asked, and feedback as written. Requests from signed-in accounts are one per account; anonymous ones are counted separately because they cannot be de-duplicated.