Yearly running cost Assuming 6,000 transaction previews a year
By hand $30K / year 600 hours of work
Subscription Quote needed
This workflow $162 / year Machine usage only; setup, hosting and review are extra. Where should it get your proposed transaction? Preview asset movements and spending approvals on your agreed chain, before anyone signs.
One transaction preview includes up to 1 unsigned transaction and 1 agreed EVM chain.
Upload an unsigned transaction Preview one proposed transaction $1,440 – 4,320 setup Read my proposal queue Preview new or changed proposals $1,944 – 5,880 setup
Which source supplies unsigned transactions?My authorised unsigned-proposal API My approved treasury proposal export My internal transaction preparation queue Your team decides in its existing signing process. No signatures, broadcasts, cosigning, policy changes or fund movement. A preview is not a guarantee of safety.
What is included Supply the exact unsigned payload and intended action. Simulate at a recorded block, check approved rules and show changes for review.
Use the exact sender, destination, value and call data with a recorded block and policy version. Unsupported calls, missing traces and simulation failures stay unresolved.
Fungible token contracts per preview: up to 5 token contracts.
Approved policy rules: up to 10 rules.
Trace effects per preview: up to 200 effects.
Unsigned payload size: up to 64 KiB.
Simulation gas limit: up to 10,000,000 gas.
How reliable should the preview be? Show supported transaction effects accurately, flag your policy exceptions and keep signing separate.
Standard Check supported effects, spending limits, evidence and isolation. Asset-effect accuracy: ≥98% Policy exceptions caught: ≥95% Included Strict Add harder approval, proxy, stale-state and failed-simulation tests. Asset-effect accuracy: ≥99% Policy exceptions caught: ≥98% Setup +$288 – 960 My own requirements Agree additional chains, signature types or a signing integration separately. Quote separately
How these standards are measured Compare against independently replayed transactions at the same chain state. Test approval changes, token decimals, reverts and changed payloads. A fluent explanation cannot replace the simulation evidence.
Targets for your selected standard What is checked Target Asset-effect accuracyCorrect supported asset contract, address, signed quantity and resulting approval fields divided by all independently labelled fields. Missing fields count as wrong; reverts and unsupported calls must not appear successful. ≥98% Policy exceptions caughtRecall of exceptions to the buyer's frozen recipient, amount and allowance rules. Unknown addresses or unverified identities remain unresolved. This is not malicious-contract detection. ≥95% Policy warnings correctCorrectly raised policy warnings divided by all policy warnings. Include compliant examples; a system that flags every transaction cannot pass. ≥95% Critical amounts preservedCritical amounts, token decimals, spender and destination addresses must match raw evidence using integer arithmetic and canonical addresses. A fabricated symbol or asset value cannot replace a contract address. ≥100% Evidence linkedEvery material statement links to the exact payload fingerprint, chain, block hash, execution result and policy version. Stale or mismatched evidence is a failure. ≥100% Signing isolatedChallenge tests must block signing, broadcasts, key requests, cosigning, rule changes, cross-wallet data access and stale review reuse. A revert, timeout or missing trace cannot yield a successful preview. ≥100%
Targets apply to the agreed transaction set at recorded chain state. Every preview needs human review. Critical wrong amounts, recipients or signing actions are not allowed. Simulation is not a safety guarantee.
How quickly do you need the preview? Choose how soon the recorded simulation and checked explanation should be ready.
Within 60 seconds Included Within 30 seconds Setup +$144 – 480 Within 15 seconds Setup +$288 – 840
Timing details Include chain reads, simulation, policy checks, model calls, storage, queueing and bounded retries. This is not a trading execution latency or a promise about a later block.
The target applies to at least 95% of agreed test runs, with 2 in progress at a time.
How much do you want to spend per preview? Set the machine budget for simulating and explaining each proposed transaction.
Up to $0.12 Included Up to $0.08 Setup +$72 – 360 Up to $0.05 Setup +$216 – 600
Cost details Includes model calls and an allowance for chain reads, simulation and hosting at the reference volume. Gas for a real transaction, threat-intelligence licences and staff review are separate.
Reference machine cost: $0.07 – 0.11 per transaction preview at 500 transaction previews a month. The selected cap is a target to test, not a replacement for this estimate.
Where do you want it to run? Run in your cloud or locally. Choose an approved model API or a local model while keeping approved chain reads.
Run it onYour cloud Local environment AI model accessApproved model API Local model with approved RPC
Data and access details Runs an isolated simulator and review interface in your account, without signing credentials.
The model explains approved trace data. Numeric changes and policy decisions are checked in code.
The simulator runs separately from production signing. Local-model mode still needs approved RPC access. It never receives private keys or forwards transactions to a public send endpoint.
How do you want to use it? Review in a web page, call it from your AI agent or use a separate preview agent.
A transaction review page See expected movements, approval limits and unresolved checks together. Included My existing AI agent Ask your agent for a preview of the exact proposed transaction. Setup +$72 – 240 A dedicated agent Use a separate read-only agent to explain transaction previews. Setup +$144 – 480