Custom keyword & regex blur lists — hide anything by pattern

Auto-detect catches standard PII patterns — emails, cards, tokens. But what about your internal project codes, client names, proprietary URLs, or custom identifiers? That's where keyword and regex blur lists come in.

What are blur lists?

Blur lists are user-defined patterns that BlurGuard checks against every visible text node on the page. When a match is found, the matching text is blurred — just like auto-detected PII, but using rules you define.

Two types of patterns are supported:

• Keywords — plain text, case-insensitive substring matching
• Regex — full regular expression syntax for pattern matching

Enabling blur lists

Toggle the etics KEY button on the toolbar, or flip the Keywords toggle in the popup. Once enabled, BlurGuard scans the page for all items in your blur list.

Click the ▾ caret next to KEY to open the keyword management dropdown. Here you can add and remove items.

Adding keywords

Keywords are simple plain-text strings. They match as case-insensitive substrings — if the keyword appears anywhere in a text node, that text is blurred.

Examples

• Acme Corp — blurs every instance of "Acme Corp" (also "acme corp", "ACME CORP")
• CONFIDENTIAL — blurs any text containing the word "confidential"
• $12,400 — blurs the specific dollar amount
• 192.168.1 — blurs any visible internal IP address starting with this prefix

Best practices for keywords

• Use specific phrases rather than single common words — "Project Phoenix" rather than "Phoenix"
• Include common variations — add both "Q3 Revenue" and "Q3 2026 Revenue" as separate entries
• Review your list periodically — remove items that are no longer sensitive

Adding regex patterns

Regex patterns use the standard /pattern/flags syntax. They match against the full text content of each text node.

Syntax

Type the pattern with surrounding slashes and optional flags:

• /ENG-\d+/ — matches "ENG-1422", "ENG-318", etc.
• /INV-\d{4}/ — matches exactly 4-digit invoice numbers like "INV-2214"
• /\b\d{3}-\d{2}-\d{4}\b/ — matches US SSN format
• /https?:\/\/internal\.acme\.com[^\s]*/ — matches internal URLs
• /[a-z]+@[a-z]+\.internal/ — matches internal email addresses

Regex flags

• i — case-insensitive (e.g., /acme corp/i)
• g — global (find all matches within a text node)
• m — multiline (match across line boundaries)

Managing blur lists in Settings

For a more spacious editing experience, open Settings (click the ⚙ icon on the toolbar or "Open settings" in the popup). Under the Blocklist section you'll find:

• A toggle to enable/disable the entire list
• A monospace textarea where you can add, edit, and remove items — one per line
• Keywords and regex patterns can be mixed freely

Changes in Settings are reflected immediately on the page — no need to reload.

Priority: keywords over PII

Keyword/regex rules have higher priority than PII auto-detection. If a keyword matches a piece of text, it's blurred regardless of the PII confidence threshold. This means you can use keywords to catch things that auto-detect might miss or dismiss.

Practical examples

Scenario 1: SaaS dashboard demo

You're demoing your product to a prospect and the dashboard shows other clients' names. Add keywords for each client name you want to hide:

Acme Corp, Globex Inc, Initech LLC, Umbrella Corp

Scenario 2: Developer tools walkthrough

You're recording a tutorial about your internal tools. Use regex to blur all internal identifiers:

/ENG-\d+/, /OPS-\d+/, /https?:\/\/dev\.internal\.com[^\s]*/

Scenario 3: Financial report

Sharing a financial report on a call. Add specific figures and internal references:

$2.4M, Q3-2026-FORECAST, / Revenue: \$[\d,]+/

Scenario 4: Customer support

Screen-sharing a support ticket with customer data visible. Combine auto-detect (for emails and phones) with keywords for internal notes:

INTERNAL NOTE, DO NOT SHARE, / escalat\w+/i

Performance considerations

Blur lists are evaluated alongside auto-detect on every page scan (initial + MutationObserver-triggered). For most lists (under 100 entries), there's no measurable performance impact. For very large lists (hundreds of regex patterns), you might notice a slight delay on heavily dynamic pages.

To optimize: use specific patterns rather than broad ones, and prefer keyword matching over complex regex when possible. A simple substring match is faster than a regex evaluation.

Keyword lists are your custom layer on top of auto-detect. While auto-detect catches known PII patterns, your blur list catches everything else that's unique to your workflow and data.