Node.js SDK /

Installation

$ npm install euredact

Upgrading from 0.6.0. Four changes in 0.7.0 can affect your code:

• Every credential is SECRET; CREDENTIAL is now a legacy alias that resolves to it. The Irish PPSN, the Austrian SVNr and the UK National Insurance number are SSN, not NATIONAL_ID.
• A span the cloud model found reports confidence null instead of "high"; filtering on "high" keeps checksum-backed rules detections only.
• A daily-quota 429 raises QuotaExceededError at once instead of after every retry.
• A cloud base URL that is not https is refused, except for localhost.

Which option do I need?

Start here when you know the task but not the option name. Every name is the camelCase form of the Python one.

redact()

Main entry point. Redact PII from a text string. Returns a RedactResult with the cleaned text and a list of detections.

function redact(
  text: string,
  options?: RedactOptions
): RedactResult

Return value

A RedactResult object with:

Detection

Each detection in the array contains:

CountryEvidence

One signal supporting a country attribution. New in 0.3.2.

DocumentContext

New in 0.3.2. Country inference reads signals out of the text, so a document processed in chunks loses accuracy at exactly the chunks that carry no signal of their own — a page of bare reference numbers between two pages full of Dutch addresses. A DocumentContext carries the evidence across chunks so the whole document is scored as one.

import { DocumentContext, redact } from "euredact";

const ctx = new DocumentContext();
let offset = 0;
for (const chunk of chunks) {
  const result = redact(chunk, { context: ctx, chunkOffset: offset });
  offset += chunk.length;
  console.log(result.redactedText);
}

// Every signal the document produced, spans pointing into the full text.
console.log(ctx.evidence);

Pass chunkOffset or the evidence spans will all point into the start of the text. Caching is disabled while a context is in use, since the result depends on the document rather than on the chunk alone.

In 0.3.3 evidence became a getter, not a method — it was a method alongside size, a getter, on the same class. Read it as ctx.evidence; ctx.evidence() now throws. The Python SDK keeps ctx.evidence(), following its own idiom.

Reuse a context only across chunks of the same document. A context shared between unrelated documents mixes their countries, which cannot cause a miss — a context influences scoring only, never which spans are found — but can attribute a value to the wrong national scheme.

Example

import { redact } from "euredact";

const result = redact(
  "Mijn BSN is 111222333 en IBAN NL91ABNA0417164300.",
  { countries: ["NL"] }
);

console.log(result.redactedText);
// "Mijn BSN is [NATIONAL_ID] en IBAN [BANK_ACCOUNT]."

console.log(result.detections);

countries takes an array. Since 0.3.3 a bare string throws a TypeError, on every entry point, in both SDKs. Before that, countries: "NL" was walked character by character into the codes "N" and "L"; neither resolves, so the call declared nothing and every detection came back with outOfScope true while the redacted text still looked correct. A pipeline filtering on that field — which these docs tell you to do — kept none of them. A wrong country code still only warns; a wrong type has no correct reading to fall back on.

redactBatch()

Process multiple texts efficiently. Loads country configs once. Same options as redact(), returns an array of RedactResult.

function redactBatch(
  texts: string[],
  options?: RedactOptions
): RedactResult[]

Example

import { redactBatch } from "euredact";

const results = redactBatch([
  "BSN 111222333",
  "IBAN DE89370400440532013000",
]);

results.forEach((r) => {
  console.log(r.redactedText);
});

availableCountries()

Returns an array of supported ISO country codes.

import { availableCountries } from "euredact";

console.log(availableCountries());  // ["AT", "BE", "BG", ...]

restore()

function restore(text: string, tokens: Record<string, string>): string

Put the original values back into text that derives from a tokenize: trueresult — typically a model’s reply to the tokenized prompt. tokens is result.tokens. Every occurrence is replaced, including a token a model has glued to other characters (EMAIL_P4RTs) — leaving a token behind is the worse failure.

Secret Detection

euRedact automatically detects secrets and API keys using two strategies: known-prefix patterns for popular services (AWS, GitHub, Stripe, OpenAI, Slack, JWT, SendGrid) and an entropy-based fallback that catches generic secrets near context keywords like api_key, token, and secret. Detected secrets are labeled as [SECRET].

import { redact } from "euredact";

const result = redact(
  "My API key is sk-proj-abc123def456ghi789"
);

console.log(result.redactedText);
// "My API key is [SECRET]"

Performance