● Offline copy

Redactor

Your offline copy. Remove client names, SSNs, and account numbers from your notes while keeping everything the tax law needs, all on this machine.

🔒 No server · no external scripts · nothing leaves this browser
⚠️

This is a safety floor, not a guarantee

Automated redaction misses things, especially names, organizations, and addresses, which are matched by best-effort pattern recognition. You must read every result before it leaves this machine. This tool reduces the chance of leaking client PII; it does not replace your review, and it does not, by itself, make the data anonymous (amounts, dates, and patterns can still identify a client). For real client data, you remain the reviewer of record.

📋
Redacting is not a license to paste client data into AI. Even a clean result is still confidential client information. Whether you can put it into a given AI tool depends on that tool's data terms (no training on your data, a business agreement, inside your WISP) and on IRC §7216 if it feeds a tax return. This tool lowers identifiability; it does not authorize the disclosure. See the §7216 decision framework (opens the live Lab site).
How this works, in plain English

You paste text. The tool reads it on your own computer and looks for things that identify a person or business: names, Social Security and tax ID numbers, account numbers, emails, phone numbers, and addresses. It swaps each one for a neutral placeholder like [Person_1] or [SSN_1], and keeps a small private list (the "local key") matching each placeholder back to the real value. That list never leaves your machine.

You take the placeholder version to your firm-approved AI tool, get your answer, then paste the answer back here. The tool uses your key to put the real names back in, right in this browser.

What it catches reliably: structured items with a clear pattern, like SSNs, EINs, emails, phone numbers, account numbers, and labeled ID / license / passport / policy numbers.
What it can miss: names, business names, and addresses are educated guesses, so it will occasionally miss one or redact too much. That is why you read the result yourself before sharing.
What it does not do: it does not make the amounts and dates anonymous, and it does not decide whether your AI tool is allowed to receive the data. Those are still your calls.

1

Paste the text you want to scrub

Paste a client note, a statement, anything with identifying details. It is processed right here in your browser. Turn WiFi off if you want to prove it to yourself.
Drop a .txt file here, or paste above. (For .pdf / .docx, use the Python tool, described at the bottom of this file.)
2

Here's the de-identified text

Read it carefully, then copy it into your firm-approved AI tool. The client's identity stays with you.
✓ De-identified, review before sharing
👀 Eyeball this before you share it. Confirm every real name, address, and account number became a [token]. If you see anything that's still identifying, it slipped through, so fix it by hand.
See exactly what was removed
🔑 Your local key (stays on this machine)

This key maps each token back to the real value. It lives only in this browser tab and never goes anywhere. Keep it with the client file. It's what you use below to bring the AI's answers back, swapping tokens like [Person_1] to the real names.

TokenTypeOriginal value

Bring the answer back

When the AI replies with tokens like [Person_1], paste its answer here. Using your local key, we'll swap the real names back in, right here in this browser. Nothing is uploaded.
Redact some text above first, then the local key is ready here.
Keeps tax facts (amounts, %, dates, states, entity type). A safety floor, not a guarantee. Eyeball the result before sharing.
Need PDF/Word support, batch files, or stronger name detection? Use the Python tool (redactor.py) shipped alongside this file. See its README.
From The AI Lab for Accountants · an educational resource, not legal or tax advice.