bitpro가 원문에서 글자만 모은 내용이에요. 표·그림·수식은 빠지거나 깨질 수 있어요. 전체 54쪽 중 4쪽
query scoped proofs (for example income above a threshold) without holding bank statements
Burnt (the company) sells verification to landlords, lenders, insurers, and other buyers who today pay traditional data brokers. Burnt's wins are **revenue**; Verona's wins are **category and network effects**—the surfaces reinforce each other.
## For developers
Start with the verification stack, then layer product logic:
| Source | Module | Developer entry |
| --------------- | ---------------- | ------------------------------------------------------------------------------- |
| Websites / APIs | zkTLS | [Internet Verification](/en/build-on-verona/verification/internet-verification) |
| Email | zkEmail | [Email Verification](/en/build-on-verona/verification/email-verification) |
| Identity | zkPassport | Truth Engine surface; contact the team for integration paths |
| Mobile apps | App Attestations | [Mobile Verification](/en/build-on-verona/verification/app-attestations) |
Publishing custom verification keys: [Publishing ZK Verification Keys](/en/build-on-verona/verification/publishing-zk-verification-keys).
Burnt Verified product APIs and issuer flows will be documented here as they ship; the modules above are the current integration path.
{% hint style="info" %}
Do not name unannounced data-broker or operator partnerships in public docs until cleared by marketing and legal.
{% endhint %}
## Related docs
* [Truth Engine](/en/about-verona/verification-infrastructure)
* [Use Cases](/en/about-verona/overview/use-cases)
* [Ero (EarnOS)](/en/about-verona/surfaces/earnos)
# Verona Litepaper
How Verona makes AI intelligent: verified facts any agent can act on.
Verona is the **intelligence layer for AI**. Prove a fact once, at the source, from a website, an email, a passport, an app, and the proof is yours. Any agent can reuse it without seeing the data underneath, and every contract is checked against a machine-verified spec. An internet of agents only works if the information holds up. Verona is the part that holds up.
## Read the lite paper
Choose your language:
| Language | Description |
| --------------------------------------------- | ----------------------------------- |
| [**English**](/en/about-verona/lite-paper/en) | Full lite paper — June 2026 |
| [**简体中文**](/en/about-verona/lite-paper/zh-cn) | 完整 Lite Paper — 2026 年 6 月 |
| [**日本語**](/en/about-verona/lite-paper/ja) | ライトペーパー全文 — 2026 年 6 月 |
| [**한국어**](/en/about-verona/lite-paper/ko) | 라이트페이퍼 전문 — 2026년 6월 |
| [**Español**](/en/about-verona/lite-paper/es) | Lite paper completo — junio de 2026 |
{% hint style="info" %}
This document is for general information only. It is not investment advice, and it is not an offer to sell or a solicitation to buy any token or other instrument.
{% endhint %}
# English
## Making AI intelligent
The intelligence layer for AI. Verified facts about the real world, owned by the user, reusable by any agent.
June 2026
***
## Abstract
Verona is the intelligence layer for AI. It makes user-verified data portable, private, and programmable, so any agent can transact on information the user actually owns.
Today an agent can talk but it cannot act, because the one thing it needs, a verified fact it is allowed to use, does not exist in a form it can reach. Verona makes that fact exist. A fact is verified once at its source, the user owns it, and any agent the user authorizes reuses it without the underlying data ever being exposed. The network combines verification at the source, trusted execution, private settlement, formal verification, and programmable payments in stablecoin. Verify once, reuse everywhere, expose nothing.
## 1 Introduction
### 1.1 The gap
AI can write, plan, and hold a conversation. It cannot act on what is true about you. Ask an agent to renew a prescription, apply for an apartment, or refinance a loan, and it stops at the first step that needs a verified fact about your income, your identity, or your eligibility. The model is fluent and the task dies, because fluency is not the same as knowing something is true. Intelligence is the ability to act on what is true, and an agent that cannot do that is not yet intelligent.
### 1.2 There is no verified data to act on
Verification happens constantly: identity checks, credit pulls, income screens. Every one of them is thrown away the moment it is used, and the next service starts over. The person the data describes owns none of it. An agent looking for a fact it is allowed to use finds nothing, because the whole stack was built for a human who could be asked to upload one more document. Agents do not upload documents. This is the gap Verona closes.
### 1.3 The unlock
Verona verifies a fact once, hands ownership to the user, and lets any agent the user authorizes reuse it. The underlying data stays where it is and is shown to no one. A single verified fact is useful; a network of them, owned by the people they describe and reachable by every agent, is a new layer of the internet. Verona is built on four ideas:
* Verified facts as a primitive - A verification becomes a durable asset the user owns, not a one-time check.
* Private settlement - Facts and the actions taken on them settle on one chain with hybrid public and private state.
* Formal verification - The chain is checked against machine-checked specifications, so an agent can rely on it.
* Programmable payments - Agents settle value in stablecoin automatically as part of an action.
## 2 Core architecture
Verona is one network. Real-world sources feed it, five capabilities make it work, and users, agents, enterprises, and the applications built on top all draw from the same set of verified facts.

### 2.1 Verification at the source
Verona never asks you to hand over your data. It produces a proof from the system that already holds the fact, and the raw data never moves. Each proof is matched to where the fact comes from: zkTLS for a web session, zkEmail for email, zkPassport for a passport, an app or enclave attestation for an application, a photo attestation for proof of a real and present person, and zkDCap, an on-chain zero-knowledge proof of a hardware attestation. Whoever consumes the proof checks it independently. Proofs and attestations trace back to their source, all smart contract and Verona source code is open and verifiable, and no part of the stack requires unverifiable trust in Verona. A recipient can audit both the provenance of the data and the validity of any selective disclosure. Verification at the source is live today against real payroll providers, and it returns a proof of income and employment in seconds. No document is stored.

### 2.2 Facts you own and reuse
A verified fact is encrypted to the user and held on the network. When the user authorizes a recipient (an agent, a service, or another enclave), the network re-encrypts the fact to that recipient and a trusted execution environment transforms it into exactly the answer the recipient needs. The recipient gets an answer and a proof, never the data. Nothing is exposed in transit, the user never has to hold a key, and one verification works for every recipient the user authorizes. Work done once is never repeated.
Stored data can only be transformed into the answer a recipient needs, which leaves no room for a false or fabricated claim. Key management runs inside hardware enclaves on verifiable, reproducible builds. Key handling is the highest-friction, highest-risk part of any blockchain tool, and here it disappears from the user's view entirely.

### 2.3 Private settlement
Facts and the actions taken on them settle on a single chain with hybrid public and private state.

What should be public is public. What should stay confidential (balances, terms, the content of a verified fact) stays confidential on the same chain. Because keys are held in hardware enclaves and the provenance stack is open, private data can produce public outputs that others can trust and act on. There is no second system bolted on for privacy. It is one chain.
### 2.4 Formal verification
Verona is built to be verified, not only audited. Every contract is checked against a machine-checked specification, and the specifications are produced through an adversarial process.

Several frontier models draft them while others attack the drafts, and a specification is accepted when it survives that scrutiny. Models that simply agree can share the same mistakes, which is exactly what the adversarial step exists to catch. This is the trust model the chain is built to, and it is what lets an agent treat the chain's behavior as something it can depend on rather than something it has to hope about.
### 2.5 Programmable payments
Native payment rails let an agent pay for what it does at the moment it does it, denominated in the stablecoin, without a person stepping in to approve each transfer. Verification, action, and settlement happen in one flow, not three disconnected systems.
## 3 The network
Users turn real-world and digital facts into reusable proofs. Builders build applications on those proofs, enterprises verify their own users and workflows, and agents act without having to trust each other or handle raw sensitive data. Every party draws on the same set of verified facts, and that is what turns a product into a network.
Verona compounds because every participant adds value to the same pool of verified facts - users add supply, applications and enterprises and agents create demand, builders expand where facts can be used, source systems increase what can be proven, and payments and settlement turn that activity into an economy.

* Users supply permissioned verified facts about themselves. They verify once, own the result, and decide which applications, enterprises, agents, or enclaves can use it. In return they get portability, privacy, and reuse, and eventually a say in how the facts they create are monetized.
* Applications create demand for those facts. They use Verona to onboard users, check eligibility, cut fraud, and complete workflows without collecting or storing raw sensitive data. Every application gives users one more reason to verify once and reuse elsewhere.
* Builders widen the network. They ship applications, agent runtimes, workflows, and integrations on top of truth primitives and skip the verification, consent, and payment plumbing they would otherwise rebuild on their own.
* Enterprises participate in two ways. They come first to verify their own users, customers, workers, transactions, credentials, and eligibility. As verified facts accumulate they stay to consume and compute over permissioned data, because verification is the first use case and verified intelligence is the larger one.
* Agents put verified facts to work, for a person, for an enterprise, or with each other.
* Source systems and verifiers expand what the network can know. Payroll providers, identity systems, banks, email, passports, apps, devices, commerce platforms, and healthcare systems each produce a proof matched to the fact being verified, and every new source makes the facts on the network broader and better.
* Token and network participants run the economic layer underneath, providing settlement, security, and liquidity for everything above it.
## 4 What it powers
These are surfaces onto one network, not separate products.

### 4.1 Attribute verification
The first surface is attribute verification. Verona can verify any attribute a user holds on a website behind a login - income and employment, spending, activity, loyalty tier, account tenure, the kind of data that normally sits siloed with no way to share it. The user logs in, and the network returns a cryptographic proof that the attribute meets a stated condition. No document is uploaded and no raw data is stored. It is faster, cheaper, and more reliable than the document-upload process it replaces, and it is live today with paying businesses. Income and employment verification for the property sector is one example. The same approach extends to lending, insurance, and any business that needs to act on a verified attribute rather than a self-reported one, with each proof reused across recipients.
### 4.2 Agents that act on verified facts
The second surface is agents, and they participate in more than one way.
* User agents act for a person, carrying that person's verified facts into a task without the user re-proving anything.
* Enterprise agents act only on verified inputs, so an automated decision rests on a proof rather than a claim.
* Agent to agent, autonomous systems transact by trusting each other's proofs instead of trusting each other, with no human in the middle to vouch.
Each case needs the same thing. A verified fact the agent can act on, and each reuses verifications already on the network instead of starting from nothing.
### 4.3 The verified context layer for builders

The third surface is builders. A single line of configuration gives any agent runtime that speaks the standard protocol access to verified, user-consented facts. Builders start from truth primitives and skip rebuilding verification, compliance, payments, and data ingestion from scratch. Verona sits between identity and tools as the layer that supplies verified context, and the proofs it returns are reusable across every other service on the network.
## 5 Why it compounds
A normal verification product starts over every time: a user proves something, an application checks it, the result dies inside that workflow, and nothing accumulates.
Verona works differently. Every verified fact becomes a reusable primitive controlled by the user, which gives applications and agents more trusted context to act on. New applications give users more reason to verify. New enterprise workflows create demand for facts that are accurate, permissioned, and source-backed. New sources widen what the network can prove.

Supply and demand reinforce each other, and the network gets stronger with use. A model that verifies and discards resets with every check and accumulates nothing, so the lead widens over time.
The moat is the growing network of reusable verified facts, owned by the people they describe and reachable by any authorized application, enterprise, or agent. No single proof, customer, or application could be copied to replicate it.
## 6 Where it leads
Verona starts by verifying facts for applications. That is the first wedge: faster income checks, employment verification, eligibility checks, credential proofs, and other workflows where businesses need to act on facts instead of claims.
As those facts accumulate, Verona becomes more than a verification layer. It becomes a permissioned intelligence layer, where enterprises and agents compute over source-backed, user-consented facts to make better decisions without exposing the underlying data. A lender can reason over verified income. An insurer can reason over verified eligibility or events. A marketplace can reason over verified reputation or transaction history. And a strategy team can finally understand market behavior from permissioned facts rather than stale panels and inferred data.
Legacy market intelligence is built on proxies: panels, surveys, brokers, scraped signals, stale datasets, and inferred behavior. Verona replaces those proxies with permissioned computation over verified facts. The user stays in control, the raw data stays private, and the output is something an application, enterprise, or agent can trust.
This is the larger direction of the network: verified data as economic infrastructure, and Verona as the place applications, enterprises, and agents go when they need to know something is true.
## 7 Token
Verona has one network token, VERONA, and one stable settlement unit, a native stablecoin. They do different jobs.
Verona is the native network token powering the intelligence layer for AI. It pays for execution under the hood, coordinates access and security, and accrues value from verified activity (i.e. revenue) on the network. The native stablecoin is the unit users, applications, agents, and enterprises use to pay, receive, and settle in a stable denomination.
Applications, agents, and enterprises can interact through familiar rails like fiat or stablecoin. Underneath that experience, verified activity creates network revenue, and network revenue creates demand for VERONA, which in turn increases network security.
## 8 Conclusion
AI becomes intelligent when it can act on what is true. That requires verified facts that are owned by the people they describe, reusable by any agent, and exposed to no one. Verona is the network where those facts live and the layer the agentic economy is built on. Verify once, reuse everywhere, expose nothing. Making AI intelligent.
## References
D. Boneh and M. Franklin. Identity-Based Encryption from the Weil Pairing. CRYPTO 2001.
DomainKeys Identified Mail (DKIM) Signatures. RFC 6376.
The Transport Layer Security (TLS) Protocol Version 1.3. RFC 8446.
Intel Trust Domain Extensions (Intel TDX). Intel Corporation.
***
This document is for general information only. It is not investment advice, and it is not an offer to sell or a solicitation to buy any token or other instrument. It should not be relied on for legal, tax, or accounting decisions. Verona is under active development, and any feature described here may change. Statements about the future are subject to change without notice. The stablecoin is a regulated, fully reserved digital dollar; any reserve-generated returns are discretionary and depend on reserve performance.
# 简体中文
## 让 AI 具备智能
面向 AI 的智能层。关于现实世界的已验证事实,由用户拥有,可被任何智能体复用。
2026 年 6 月
***
## 摘要
Verona 是面向 AI 的智能层。它使用户已验证的数据具备可携带性、隐私性和可编程性,让任何智能体都能基于用户真正拥有的信息进行交易。
今天,智能体可以对话,却无法行动,因为它所需要的关键要素,即一个被允许使用的已验证事实,并不存在于它能够访问的形态中。Verona 让这个事实真正存在。事实在源头被验证一次,由用户拥有;用户授权的任何智能体都可以复用该事实,而底层数据永远不会被暴露。该网络结合了源头验证、可信执行、私密结算、形式化验证,以及以稳定币进行的可编程支付。验证一次,处处复用,绝不暴露。
## 1 引言
### 1.1 缺口
AI 可以写作、规划并进行对话。它却不能基于关于你的真实事实采取行动。让一个智能体续开处方、申请公寓或为贷款再融资,它会在第一个需要关于你收入、프로젝트로 돌아가기