The Verdict
Put a claim to an impartial AI judge that reads the world and rules, on-chain.
Pay three independent AIs a few cents to check one answer, and get a signed receipt of agreement.
Try it live, freePriya gets a contract summary from one model. She sends it to Second Look, three verifier agents paid via x402 score it, and she gets an onchain attestation she can show her client.
Verifier panel of 3 models, x402 payments, EAS attestation, shareable receipt page
Reshape it entirely: new features, your flows, your brand end to end, integrations you need. We scope it with you.
Multi-model verification with signed receipts already exists; per-answer micropayments are the remaining twist.
Synod (ETHGlobal) has multiple models vote and post a signed receipt onchain, MAMV produces vote receipts with onchain anchors, and Hive SiGR sells signed multi-model panel receipts.
Checked with web searches in October 2026. A search is not proof that nothing similar exists, and the market moves fast, so we recheck before every build.
Onchain AI apps; eligibility unknown
AI tracks; varies per event
The complete Second Look app: send one AI answer and its source to a panel of three independent verifier models paid a few cents each over x402, and get an onchain receipt of agreement you can show your client.
Every buyer gets the same finished app. Basic branding and copy changes come through a quick form.
Your admin side: revenue, paywall editor, users, rules and the agent log.
Referral and growth mechanics, switched on when you are ready to grow.
Pay as you go users pay verifier cost plus a small margin per check; professionals and firms pay monthly for workflow features.
Margin per pay as you go check set by the owner (default $0.02), and an optional listing share from verifier operators.
Prices are yours to set. Illustrative, not a promise of income.
Pay three independent AIs a few cents to check one answer, and get a signed receipt of agreement.
Problem: Professionals send AI written summaries to clients with their name on them, but one model can be confidently wrong and there is no record of how it was checked.
Solution: A panel of three verifiers from different labs is paid per check over x402, votes are aggregated by the user's rule, and an attestation records the document hash and votes.
Why now: x402 makes paying several models per request trivial, and cheap attestations make a verifiable receipt practical for every document.
Market: Freelance consultants, translators, paralegals and small firms that send AI assisted work to clients.
The ask: Pilot with 50 professionals and three model operators, plus ecosystem grant support.
Funding is never guaranteed. Check each program's rules.
Many hackathons require code written at the event. See the Regatta Boardfor each race's rule.
Custom work from $1,000: your own verifier list, a firm branded receipt portal, document tool add ins, or audit exports. Scoped on a call.
More and more work that goes out under a person's name is drafted by one AI model. One model can be confidently wrong, and the person sending it has no cheap, quick way to get a second look or to show a client it was checked.
Multi-model verification with signed receipts already exists: Synod (an ETHGlobal project) has several models vote and posts a signed receipt onchain, MAMV produces vote receipts with onchain anchors, and Hive SiGR sells signed multi-model panel receipts. Research verdict: exists, low novelty. The remaining twist here is per-answer micropayments to independent verifiers over x402, a budget and panel picker for non-technical professionals, and claim by claim checks that must cite the source.
The hash of the document being checked, the answer or summary under review, each verifier's vote and clause citations, x402 payment receipts and the resulting EAS attestation. Document text is sent only to the chosen verifiers and is never written onchain.
Make it normal and cheap to have AI work checked by independent eyes before it goes out with your name on it.
Less than a coffee A three model check costs a few cents, so people run it on everything, not just the scary documents.
A receipt you can show Clients get a link with the seal: this exact file, these verifiers, these votes.
The Friday send-out Nothing leaves the outbox unsealed. A streak of clean Fridays is a quiet point of pride.
Upload and split: The app splits the answer into claims and pairs each with the source passages it should match. Text is hashed locally (SHA-256) before anything is sent.
Verifier panel: Three models from different providers via OpenRouter or direct APIs, for example a frontier reasoning model, a mid tier model and a small fast model. Each must cite a clause for every claim it accepts or rejects.
x402 payments: Each verifier endpoint quotes a price with HTTP 402; the app pays in USDC on Base only if the quote is within the user's per check budget, and never retries at a higher price silently.
Aggregation: Votes are combined with the user's rule (2 of 3 or 3 of 3). Disagreements are shown claim by claim with the verifier's reason.
Attestation: An EAS attestation on Base records the document hash, verifier ids, votes, rule and payment references. The schema is public so anyone can check a receipt.
Share page: A receipt page shows the seal and votes. Verifier notes are visible only to people with the link; the document itself is never published.
Per check: model tokens for a 2 page summary against a 10 page contract are roughly 0.5 to 5 cents depending on the models chosen; an EAS attestation on Base is usually well under 1 cent of gas. A user running 100 checks a month spends about 3 to 8 USD on verifiers and gas.
Every receipt link sent to a client shows how the check was done and invites them to check their own documents. Verifier operators can list their own models and earn per check, which brings new supply and their audiences.
False comfort: Three models can share a blind spot. The receipt says what was checked, not that the document is correct, and the app says so on every seal.
Confidentiality: Contracts are sensitive. Send text only to providers with no-training terms, let users pick self hosted verifiers, and keep only hashes onchain.
Correlated verifiers: Models from the same lab or trained on similar data agree too easily. Warn when a panel is not diverse and track each verifier's catch rate.
Price games: A verifier could raise prices mid request. Pay only against the quote, cap per verifier price, and log changes.
Not legal advice: The app checks consistency between a summary and its source. Professionals remain responsible for their own advice.
Costs and tools are estimates at the time of writing. General information only, not legal or financial advice.
Nine routes from one vessel. Tap a port to open its chart.
People are sending AI written contracts, reports and summaries to clients with their own name on them, and they have no way to show it was checked. Second Look pays three independent AI verifiers a few cents each through x402 and issues a signed attestation of how far they agreed. If you own it, the verifier panel, the pricing margin and every receipt carry your brand, and you earn on each check instead of paying someone else's.
Yes. The panel of three models, x402 payments, EAS attestation and receipt page are built. You choose the chain, models and branding, add your API keys and deploy from your wallet with the guide.
No, and say so clearly. Agreement lowers the chance of an obvious error but models can share the same blind spot. Position the receipt as evidence of checking, not a guarantee of truth.
Each check pays the verifier models. Set your price above your cost, for example charge 25 cents for a check that costs you 8. Watch costs weekly and cap daily spend.
Freelancers, consultants and small agencies who send AI assisted work to clients. Start with one profession you know, like contract summaries for freelancers.
Documents go to the model providers you choose. Tell users which ones, let them pick providers with suitable data terms, and warn them not to submit anything they are not allowed to share.
You charge per check (or a monthly bundle) and keep the margin above model costs. Teams pay monthly for a volume of checks plus branded receipts their clients can open.
Illustrative stages, not a forecast. Results depend on you and the market; nothing here is guaranteed.
Put a claim to an impartial AI judge that reads the world and rules, on-chain.
Invite-gated: Ritual's testnet faucet needs a Discord invite. Get the invite, then contact us and we set you up. Not sold in the normal flow yet.
Teach an AI agent your craft, and every time it earns, you earn as the teacher of record.
An AI that guards your inbox and DMs and charges strangers a fee you keep to reach you.
Give your AI agent a wallet with house rules it literally cannot break.
Your personal AI memory, owned by you, rentable to any assistant you choose.
AI labs pay you to rank their outputs, and your taste builds an onchain reputation.
A public, verifiable work history for any AI agent, built from the jobs it actually finished.