Live demo
BeginnerTry it: check member data before you paste it
This is the working version of our member-data tutorial, running right here in your browser. Paste a draft, watch the checker flag the risky bits, redact them all, then re-scan — the same "test your cleaned text before you paste" step the tutorial ends on. Nothing leaves this page.
What you just ran
Seven pattern detectors, each looking for one category the tutorial
flags: email addresses, US phone numbers, Social Security numbers,
street addresses, credit-card-like numbers, dates of birth, and
member/customer ID patterns. Every hit is highlighted with its category,
Redact all swaps each one for a
[REDACTED: CATEGORY] placeholder, and Re-scan
proves the text is clean — the tutorial's final step, automated.
- Step 4, automated — the flagging pass that finds what names a person or describes money. Regex does the mechanical part; your judgment does the rest.
- Step 6, automated — every flagged value becomes a square-bracket placeholder, the way the tutorial teaches.
- Step 7, automated — re-scan confirms zero flags remain before anything goes into an AI tool.
What the regex can't do is the human part: names don't match a pattern, and three innocent details can still point to one member. That's why the tutorial ends with a colleague check — this demo handles the mechanical pass so your eyes can handle the rest. Read the tutorial.
Request the starter kit
The demo above runs a simplified version in your browser. The starter kit is the real thing: a single Python file implementing this exact detection logic for your own machine, plus the tutorial's eight steps as a printable checklist.
Want the files behind this lab, sample dataset included? Email ai@trihelix.co and we will send them your way.