Redaction

What is Redaction?

Redaction feature lets you hide sensitive portions of an HTTPS transcript from the Prover. Common use cases include removing cookies, authorization headers, or API tokens before generating a Web Proof. At the same time, everything you leave visible is still cryptographically verified for integrity.

⚠️ Warning! Unsafe byte-range redaction can introduce ambiguities and vulnerabilities. This guide explains how to use redaction safely.

To learn how to enable and configure redaction using the vlayer SDK, see the Redaction section in our JavaScript documentation.

Currently supported redaction targets:

URL path
Request headers
Response headers

Note: Redaction of the response body is currently supported only in unsafe mode, which demands special security concerns. More details will be provided in a future book update.

Security Model

Why Caution Is Needed?

In TLSN Protocol, the foundation for Web Proofs, redaction is performed over raw byte ranges. This means the transcript is treated as an unstructured byte stream, without awareness of HTTP headers, query parameters, or other protocol elements.

For example, this TLSN function redacts bytes from 2 to 4.

redact(2, 4)

Let’s take a closer look at security nuances of specific cases.

URL Redaction

Consider the following redacted URL path (* stands for redacted character):

/user?name=******************

This redacted form could correspond to multiple original inputs, such as:

/user?name=John&surname=Smith
/user?name=JohnathansLongName

Without access to the hidden portion, it is impossible to determine the original URL from the redacted version. This ambiguity arises because the redaction process operates on raw byte ranges of the same length. As a result, only the number of redacted bytes is known, while the structure of the redacted path remains unknown.

Enforcing URL Integrity

To guard against URL redaction issues, WebProof provides two verification modes. They limit the way url can be redacted.

1. Full-URL verification

Use the verify function to check the integrity of the entire, unredacted URL. This mode resolves URL redaction issues by completely disallowing URL redaction. Example:

function main(WebProof calldata webProof) {
    Web memory web = webProof.verify("example.com/user?name=John&surname=Smith")
    ...
}

2. URL prefix verification

Use the verifyWithUrlPrefix function to validate that the redacted URL starts with a known prefix. Example:

function main(WebProof calldata webProof) {
    Web memory web = webProof.verifyWithUrlPrefix("example.com/user?name=")
    ...
}

Assumes the prefix (“example.com/user?name=”) is correct.
Treats everything after that prefix as untrusted (it can be redacted).
Ensures that sensitive suffix data (e.g. user IDs) remains hidden, while protecting contract logic from tampering.

Details on how to prepare a WebProof with redacted URL can be found here.

Header Redaction

Both verify and verifyWithUrlPrefix functions handle header redactions in the same way, and no additional precautions are required to use them safely. Details on how to prepare a WebProof with redacted headers can be found here.

vlayer book