Redact financial data before you paste it into an AI
Bank statements, invoices and payment disputes are the documents people most want a model to explain and least want to hand over. This preset turns on every financial detector — card numbers, IBANs, routing numbers, account references, crypto wallet addresses — alongside names, companies and addresses.
Amounts, dates and transaction descriptions stay, because those are the parts the model needs to categorise spending, find a duplicate charge or explain a fee.
Your text
Safe to paste
Preset: every financial identifier, plus names and addresses. Amounts and dates kept.
Mapping — 0 values (only visible in this tab)
| Placeholder | Real value | Uses |
|---|
Check a few rows before you paste — this is the fastest way to spot a detector that grabbed the wrong thing. Nothing here is stored or sent anywhere.
What to look for — 30 detectors, custom terms, exceptions
People & organisations
Government & health IDs
Financial
Technical & secrets
Location
Step 3 — paste the AI’s answer here to put the real values back
The model replies using the placeholders. Paste its answer below and every
[NAME_1] becomes the real name again — using the mapping held in this tab, which
is the only copy that exists.
How it works
- Open this page — the detectors are already set for “Financial data”.
- Paste your text, or drop a .txt, .md, .csv, .json or .log file onto the panel.
- Check the highlights: every match is coloured by type, and one click keeps a type as-is.
- Copy the safe version and paste it into the AI tool of your choice.
- Paste the model’s reply back into step 3 to restore the real values.
Checksums keep the false positives down
Financial documents are full of long numbers that are not account numbers, so every financial detector here validates rather than guesses. Card numbers must pass the Luhn check and start with a real issuer prefix. IBANs must pass the ISO 13616 mod-97 test. US routing numbers must satisfy the ABA 3-7-1 weighted checksum and sit next to a routing label. The effect is that reference numbers, invoice IDs and transaction codes survive the scrub, which keeps the statement legible.
Amounts are never touched. A model cannot categorise your spending without them, and an amount on its own identifies nobody.
What a statement still says about you
Redacting the account number does not anonymise a bank statement. The merchant list is a portrait: the gym, the pharmacy, the parish, the town the coffee shop is in, the salary that arrives on the 15th from a named employer. Anyone reading it can infer where you live, work, worship and receive care.
If that matters for your use, cut the statement down before you paste — one month, the categories you are asking about, the merchant names generalised. Use custom terms for the employer and any merchant you would rather not name. The scrub protects the identifiers; only editing protects the pattern.
Tips
- Add your employer and bank to custom terms if the question does not depend on them.
- Crypto addresses are permanently public on-chain — replacing one here protects the link between it and you, not the address itself.
- For a dispute letter, use fake values so the model drafts natural prose, then restore the real numbers in step 3 before sending.
Detectors are pattern-based and imperfect: read the highlighted output before you paste it anywhere, and confirm your own confidentiality obligations against the current rules for your profession, employer or jurisdiction. Compiled 2026-07-29.
Frequently asked questions
Is the last-four of a card detected?
No, by design. Four digits are not a card number, they are the identifier businesses legitimately use to discuss a payment, and redacting them would break the conversation without protecting anyone. Full numbers are detected and replaced.
Does anything here get sent to a bank or a validation service?
No. Every check is arithmetic done in your browser: Luhn, mod-97, the ABA weighting. There is no lookup, no API and no network request of any kind while you type.
Can I restore the real numbers in the model’s reply?
Yes, with the tokens or fake-values style. Paste the reply into step 3 and each placeholder becomes the original value again. The blocks and delete styles are one-way on purpose.