Privacy before you paste

Free
0/7 steps visited

Step 3 of 7

Red, amber, green data

Read · ~6 min

Red, amber, green data

Treat this like kitchen hygiene — not perfectionism, just sensible defaults you don't have to think hard about.

Red — never paste into cloud AI

ExamplesWhy
Passwords, API keys, recovery codesInstant account-takeover risk
Bank / card details, full payment infoFinancial fraud
Customer medical or children's detailsYou may break both trust and the law
Unredacted contracts with names + figuresClient confidentiality
Someone else's private messages, forwarded in fullYou don't own that consent

The rule: if you wouldn't email it to a random help desk, don't paste it.

Amber — careful (anonymise, summarise, or wait)

ExamplesSafer approach
A real client scenario you want advice onChange the names, amounts and location; say "fictional example"
Internal business numbersRound them off; strip the identifiers
A draft policy you're unsure aboutPaste your question, not the whole employee handbook
Competitor research with their brandingDescribe the situation in your own words

The rule: ask "if this leaked, who gets hurt?" If the answer isn't "nobody," it's amber.

Green — usually fine for learning and daily drafts

ExamplesStill check facts
Public facts, general how-to questionsAI can still be wrong
Your own marketing angle (no client secrets)
Code with no secrets in it (fake keys in exercises)Use YOUR_API_KEY_HERE placeholders
Lesson exercises with the made-up shop "Example Ink Ltd"

Amber is where I live most days, honestly — the real shape of a problem with fake names bolted on. "A customer in Sheffield, deposit dispute, £200" — not their actual Instagram handle and a screenshot of the invoice.

Quick self-test before you hit Enter

  1. Whose data is this? (Mine / a client's / someone else's)
  2. Could it identify a real person?
  3. Is there a smaller paste that still gets the job done?

Continue — five minutes inside your app's settings.

Warning

Real power. Educational use only.

What we teach you to build is genuinely powerful — uncensored assistants, agents, and automations on your own hardware. In the wrong hands, that is as dangerous as malicious code in the wrong hands. We do not teach illegal, malicious, or harmful use. You are responsible for what you deploy.

See what we mean →