PureGuard – Bot & Ad-Fraud Protection

By Pure Guard

0+ active installationsVersion 6.5.0

Bot & Ad-Fraud Protection: checks every visitor, lets real people in and stops bots. Start in Monitor mode and watch before anything is stopped.

Description

PureGuard checks every visitor to your WordPress site, lets real people in and stops bots. You pick one of six modes and the plugin sets everything else for you.

Every visitor is put in one of three groups:

  • Real people — visitors we are sure about.
  • Unsure — not clearly a bot, not clearly a person.
  • Bots — confirmed bots.

The six modes

  • Off — Nothing is checked and nothing is recorded.
  • Monitor (watch only) — Every visitor is checked and shown on the Live page, and everyone is let in. New installs start here: watch for a few days before you stop anyone.
  • Protect — Confirmed bots are stopped. Everyone else is let in without noticing anything.
  • Protect + browser check — Confirmed bots are stopped, and visitors we are unsure about get a short automatic browser check that real people pass in a few seconds.
  • Strict — Only visitors we are sure about are let in. Some real people on unusual browsers will be stopped too.
  • Emergency — Only verified Google, Bing, Apple and Meta crawlers get in; every other visitor is stopped, including other search engines, uptime monitors and ad-verification crawlers. Logged-in users are never stopped, so you cannot lock yourself out. Use during an attack, then switch back.

Visitors on VPNs

If your real audience uses VPNs or proxies to get around censorship, choose “My readers use VPNs to get around censorship”. A visitor is then never stopped for their network alone. With the browser check they get the quick check instead (people pass it, bots cannot). Visitors with real bot evidence are still stopped.

Your own look on the browser check and block pages

Add your own logo, brand name, accent color, and dark or light theme. Change the block-page heading and message, show stopped visitors their IP and an incident ID, and let real people report a mistake with one click (“I am human”). Preview both pages from the settings before anything goes live.

Live page

  • Protection status — whether the site is connected, the current mode, when the last visitor was checked, and your plan usage.
  • The last 24 hours — visitors checked, let in and stopped, verified search crawlers, a chart by hour, why visitors were stopped, and the top countries.
  • Ready for Protect mode? — while you are in Monitor mode, the page tells you whether a blocking mode would have stopped real visitors or search crawlers yesterday.
  • This site’s own history — a live feed of the latest visitors, the addresses seen most as bots, visitor reports, daily totals, and a CSV export of the last 7 days.

These numbers cover visits that reach WordPress. Pages served from a page cache or CDN cache are not seen by the check. The history is kept in your own database (last 30 days).

Browser check page in several languages

The browser check page picks the visitor’s browser language by itself. English, Thai, Bahasa Indonesia, Vietnamese, and Burmese (Myanmar) are built in, and every line can be edited in the settings.

Visitor engagement (off by default)

Optional: when you turn it on, the plugin measures anonymous time on page and scroll depth for real visitors. Off by default — no measurement script is added unless you turn it on. No personal data is collected.

External services

This plugin connects to the PureGuard service (https://pureguard.io) to check visits:

  • What is sent: visitor IP address, user agent, requested URL/campaign parameters and browser-provided headers of the request being checked; your API key authenticates the calls. The Live page also asks the service for this site’s 24-hour summary, sending the site’s host name.
  • Browser verification (optional, off by default; the two “browser verification” switches under Paid campaign traffic): when on, the visitor’s own browser sends a short anonymous measurement straight to https://pureguard.io — screen size, graphics hardware name, processor count, time zone, language, whether the page became visible and whether there was real scrolling, touch or pointer movement. No cookies are set by it and no names, e-mail addresses or form contents are read. Turn it on only if your privacy notice covers it.
  • Browser check for cached pages (optional, off by default): when on, the reader’s browser sends one short request per visit straight to https://pureguard.io/page_verdict.php containing four things: the site’s host name, whether the browser shows signs of automation (and which signs), whether the page was visible, and whether the device has a touch screen; the service also sees the reader’s IP address and user agent, as any web request does. No cookie is set; the answer is kept in the browser’s session storage for ten minutes. On a site that uses the WP Consent API nothing is sent until the reader has allowed the “statistics” category. Turn it on only if your privacy notice covers it.
  • When: on protected page requests (bot check), when you press Test key, when an administrator opens the Live page, and when campaign features are turned on.
  • Why: the checks run on the PureGuard servers; the plugin only passes on the request details and applies the answer. If the service cannot be reached, visitors are let in.
  • Service terms: https://pureguard.io/terms — Privacy: https://pureguard.io/privacy

FAQ

Will this stop my real visitors?

In Off and Monitor mode, never: nobody is stopped. In Protect mode, only confirmed bots. In Protect + browser check mode, visitors we are unsure about briefly see an automatic browser check. In Strict mode, unsure visitors are stopped too. In Emergency mode everyone is stopped except logged-in users and verified Google, Bing, Apple and Meta crawlers. Start in Monitor and use the “Ready for Protect mode?” card on the Live page before you switch.

My readers use VPNs. Will they be stopped?

Under Settings → Protection → Visitors on VPNs, choose “My readers use VPNs to get around censorship”. Visitors marked only for their network are then never stopped; at most they see a quick automatic browser check that real people pass in seconds.

Do I need a PureGuard account?

Yes. Add your API key under Settings → PureGuard and press Test key. Get a key at https://pureguard.io/register.

Does it slow down my site?

A visitor’s result is remembered (1 hour for visitors let in, 5 minutes for visitors stopped), so the service is asked at most once per visitor in that time. Logged-in users are skipped. In Monitor mode the page is sent first and the check runs afterwards where the server supports it. If the service cannot be reached, visitors are let in.

My site is cached. Does the plugin still see my readers?

Only some of them. The server check runs when WordPress builds a page. A reader who is served a stored copy from a page cache or a CDN never reaches it, and on a well-cached site that is most readers. To see those readers too, turn on “Check readers in their browser” under Settings → Protection → Readers on cached pages. The check then runs in each reader’s own browser, also on cached pages, and the Live page shows the result. It stops nobody; it tells your pages whether the reader looks real. If you choose, it also passes the result to Google Ad Manager as a key-value, or holds AdSense ad requests for confirmed bots.

Where is the traffic data stored?

In a small table in your own WordPress database, cleaned up automatically after 30 days. You can turn this history off in Settings → General.

Screenshots

Changelog

6.5.0

A check that also works on cached pages.

  • New, off by default: “Check readers in their browser” (Settings → Protection → Readers on cached pages). On a site with a page cache or a CDN most readers are served a stored copy and the server check never sees them. With this on, each reader is checked in their own browser instead. It adds one small script to your pages and one short request per visit. Nobody is stopped by it. The Live page shows how many readers were checked this way and how many looked real, unsure, bots or crawlers.
  • New, off by default: send the result to Google Ad Manager as the key-value pg_ivt (ok, unsure, bot, crawler), so you can report on bot traffic or exclude it in your ad server.
  • New, off by default: do not request AdSense ads for confirmed bots. AdSense ad requests wait for the result, at most 1.5 seconds, then go ahead anyway. If a consent tool has already paused ads, that is left alone. For Google Publisher Tag or other ad code use the key-value or the developer hook. Test it on your own pages first.
  • For developers: pgVerdict.on(function (v) { ... }) gives your own scripts the result; there is also a pgverdict event on document and a data-pg-verdict attribute on the html element.
  • Unrecognised proxy or CDN: the safety valve now remembers up to 32 addresses instead of one, for a CDN that reaches your site from several. Each address still has to carry many different visitors before it counts, and exempts only itself. The admin notice groups the addresses by network range so they are easy to add under Trusted proxies.
  • If your site uses the WP Consent API, the browser check sends nothing until the reader has allowed the “statistics” category.
  • The plugin is now called “PureGuard – Bot & Ad-Fraud Protection”. Settings, menus and your saved options are unchanged.

6.4.0

Security hardening and a clearer admin area.

  • New installs start in Monitor mode (watch only). Existing sites keep their mode.
  • After activation you land on the Live page; while no API key is saved it shows three steps to connect.
  • Live page: new “Protection status” section with the last 24 hours (visitors checked, let in, stopped, verified search crawlers, chart by hour, why visitors were stopped, top countries) and a “Ready for Protect mode?” card. Results and reasons are shown in plain words. The Performance view is gone; daily totals moved here from the Stats tab, which was removed. New CSV export of the last 7 days.
  • Mode names: Medium is now “Protect”, High is “Protect + browser check”, Lockdown is “Emergency”. Your saved mode is unchanged.
  • Settings: “Trust VPN users” and “Censored audience” are one choice, “Visitors on VPNs”; paid campaign options and rarely used settings are folded away under “Paid campaign traffic” and “Advanced”; “Add my current address” button for Trusted IPs; “Disconnect” removes the saved API key.
  • White-label now also covers the WordPress admin area (menu, page titles, dashboard widget, notices, Plugins row).
  • The proxy notice appears only when one unrecognised address carries a large share of the visits and passes on many different visitor addresses.

  • Visitor address: forwarded headers (CF-Connecting-IP, X-Forwarded-For, X-Real-IP) are now believed only when the request really comes from Cloudflare, from a proxy on your own server or network, or from a proxy you list under the new Advanced → Trusted proxies setting. Otherwise the address of the direct connection is used. If your site sits behind another proxy or CDN that we do not recognise, visits through it are only watched (nobody is stopped, except in Emergency mode) and a notice on every admin screen tells you which address to add.

  • Search crawlers: a visitor that only claims to be a Google, Bing, Apple or Meta crawler is now checked against the address ranges those companies publish. Nothing is blocked by default: unverified crawlers are counted on the Live page. A crawler is never delayed by this check, and never treated as a pretender when the service has no crawler address data for its answer (counted as “not checked”). A new setting, “Block visitors that pretend to be Google, Bing, Apple or Meta crawlers”, applies the normal verdict to them once you have watched the count.
  • Emergency mode (formerly Lockdown) now also applies to the other self-declared crawlers (they used to be let in by name). Google, Bing, Apple and Meta crawlers with a verified address still get in.
  • Campaign cookie is now signed, tied to the browser and expires; a hand-made cookie no longer skips anything. Campaign “Measure only” mode is unchanged: the visit is let in after the one campaign measurement, so site protection does not stop campaign visitors while it is on.
  • Browser check: each challenge is tied to the visitor’s address and browser and can be used once. If the address or browser changes during the check, or a check is submitted twice, the visitor gets a fresh check (up to 5 in 10 minutes) instead of the block page.
  • Removed the option that reported background tabs as visible to ad tags (pgperf_force_render).
  • A cached verdict is kept at most 1 hour for a pass and 5 minutes for a block or challenge, so one bad verdict cannot stick to a shared address.
  • The campaign check uses the same circuit breaker as the site check (2-second limit), and a slow answer no longer counts as a recovery.
  • Fewer database writes on busy sites (zone status at most once per 5 minutes, log pruning in batches).
  • The render log moved to a protected sub-folder of uploads, and its endpoint is rate-limited per address.
  • The redirect allow-list for your quarantine / after-check URLs is applied only around the plugin’s own redirects.
  • A campaign token replayed from a different address is refused.
  • The API key is no longer printed into the settings page; leave the field empty to keep the stored key.
  • New installs start with the quiet log on (no per-visitor line in the PHP error log). Existing sites keep their setting.
  • Gate, block and challenge pages are marked uncacheable and cannot be framed; the verified cookie is SameSite=Lax.
  • Malformed replies from the service can no longer cause a PHP error.
  • Uninstall also removes the newer options, the log folder and leftover campaign-token transients.
  • Works with LiteSpeed’s finish-request as well as PHP-FPM’s.

6.3.6

  • The site check now tells the PureGuard service whether the page carries the browser-evidence collector, so “this visitor never ran the page’s JavaScript” is only counted on pages where the collector was actually present. Reporting accuracy only: no change to any verdict, challenge, redirect, cookie or page HTML.

6.3.5

  • Live header shows the last successful contact with the PureGuard service (refreshed at most every 5 minutes) and, for your own site, today’s checks and accept rate as the server reports them. No change to any verdict, challenge, redirect or cookie.

6.3.4

  • New option “Measure reader pages too”: the collector can also run on ordinary reader pages (off by default), so the decisions on organic readers can be measured against real render and interaction behaviour. Page HTML stays identical for every visitor; caches keep working.

6.3.3

  • Campaign pages: the collector now reports whether the visitor actually interacted with the page (first tap / key / scroll) and how long they stayed, so engagement is measured instead of guessed. One extra lightweight beacon per visit; nothing is blocked or redirected by it.

6.3.2

  • Trusted IPs: list your own servers (feed fetchers, uptime monitors, cache preloaders) so they are never checked or redirected.

6.3.1

  • Sites using Cloudflare Pseudo IPv4 (overwrite): IPv6 visitors are now checked by their real IPv6 address instead of the 240.0.0.0/4 placeholder, which every bogon list treats as invalid.

6.3.0

  • Campaign gate observe mode: the full campaign check runs on paid visits and is recorded per ad zone, but no visitor is ever blocked or redirected. Use it for the first week of a new campaign, then exclude bad zones at the ad network. Fix: the configured dump URL host is now allowed for redirects (blocked visitors were sent to the login page on sites without a separate allow-list).

6.2.9

  • Uninstall now removes every plugin option.

6.2.8

  • Optional quiet log: no per-visitor line in the PHP error log. Neutral log prefix in white-label mode.

6.2.7

  • White-label mode now also uses neutral names for the challenge form fields, cookies and script variable. Both name sets are always accepted, so switching the option never breaks a page or a visitor’s pass.

6.2.6

  • Optional white-label mode: challenge and block pages show only your site name, with styles and scripts inlined so no plugin file paths appear in the page source.

6.2.5

  • Optional reader pass: after a visitor passes the browser check, that browser is never blocked or challenged again for the configured number of days (default off, recommended 28 for news sites), even when its IP changes. Device-bound, signed, and revoked automatically on real bot evidence. Campaign traffic and Lockdown are not affected.

6.2.4

  • Censorship mode no longer softens FireHOL blocklist hits (abuse evidence, not a network reason). Optional Cloudflare-verified crawler trust via a secret header set by a Cloudflare Transform rule (off unless configured). Uptime-Kuma recognised as an uptime monitor.

6.2.3

  • Monitor mode no longer delays the page: the protection check runs after the response is sent (PHP-FPM); other hosts behave as before. Saving one settings tab no longer resets the options on the other tabs. Gate redirects send no-cache headers. Render counter honours its switch, clamps field lengths and stops at 20 MB. Verdict cache is per IP + browser (a cache time of 0 now means the 1-hour default). robots.txt, ads.txt, app-ads.txt and sellers.json are never gated.

6.2.2

  • Canonical signed visit identity: the plugin forwards the PureGuard-signed visit id (first-party pgvid cookie set from the server response, or the signed ?pgvid hand-off on the redirect) so browser evidence joins the exact decision it came from. Never minted client-side; never in page HTML (page caches stay safe).
  • The L3 browser battery is printed inline, first in the document, on campaign pages only (no external asset for a CDN or browser to hold stale).
  • Campaign URLs accept zone= as an alias of s1=; click_id and bid macros are forwarded to the PureGuard API.
  • Settings page regrouped: crawlers, VPN audiences, paid-traffic options. Publisher-site zone exemption is now a named, default-off policy.

6.2.1

  • Campaign gate now forwards the ad network’s click_id and bid macros to the PureGuard API, so campaign spend and click joins work on plugin-delivered traffic (both optional, safely validated).
  • Settings copy clarified.

6.1.0

  • FIXED: On the publisher smart-route, one visit could trigger two protection decisions; every visit now receives exactly one server-side Guard decision (faster page delivery, cleaner analytics).
  • IMPROVED: Clearer product language across the settings screens — “Deep browser detection”, “Bot quarantine URL”, “Browser protection” — describing what each control does without internal jargon.
  • IMPROVED: Account statistics wording now consistently reflects PureGuard’s own first-party intelligence.

6.0.23

  • NEW: Optional central protection policy for sites linked to a PureGuard account — the client-side browser-protection check can be managed centrally over a signed, authenticated connector channel. It fails safe: if the central service is unreachable or a response can’t be verified, browser protection stays ON. Sites that don’t use a PureGuard Connector are unaffected and keep their existing local setting.
  • SECURITY: Connector credentials are now stored encrypted at rest.
  • IMPROVED: Per-request diagnostic policy logging is emitted only when WP_DEBUG (or a dedicated debug flag) is enabled; normal production stays quiet.

6.0.22

  • IMPROVED: Campaign render counting now fires the moment the page arrives (with a follow-up once the ad slot actually fills), so pop-under landings opened in a background tab are no longer missed by the counter.

6.0.21

  • NEW: First-party render counter for campaign landing pages — a tiny same-origin beacon (immune to ad blockers) records that your money page really rendered and whether the ad slot filled. Stored only in a private file on your own site under a non-guessable name; nothing is sent anywhere. On by default for campaign visits only; switch it off in Settings.

6.0.20

  • NEW: Optional “force render” for pop-under campaign landings (OFF by default). Some ad tags wait for the tab to gain focus before rendering; this treats the freshly opened background tab as already visible so the page and its ad render on load. Applies to campaign traffic only and never touches normal visitors.

6.0.19

  • IMPROVED: The same-origin campaign gate now forwards the visitor’s real browser signals to PureGuard, so genuine humans are scored with full evidence instead of IP + user-agent alone. Fewer false quarantines of real visitors; bots gain nothing (the signals are browser-enforced).

6.0.17

  • NEW: Same-origin campaign gate (opt-in, OFF by default) — filter your ad-campaign traffic right on the landing page instead of routing it through a separate gate URL first. Blocked ad zones still apply; if PureGuard is ever unreachable the gate fails open and lets the visitor through, never breaking your page.

6.0.16

  • IMPROVED: The plugin records your workspace key once from your PureGuard account, so campaign intelligence beacons are correctly attributed to your workspace.

6.0.15

  • NEW: Optional client-side browser-signal collector for campaign landing pages (OFF by default). Collects the same in-browser bot tells as the hosted PureGuard gate — same-origin, so nothing extra is loaded from third parties — and sends them to PureGuard for scoring. The detection logic itself stays server-side.

6.0.14

  • SECURITY: The engagement beacon no longer places your API key in the page source. It now posts to your own site, which forwards it to PureGuard server-side — the key never reaches the browser.
  • NEW: The Performance view now shows real visitor engagement — average time on page, scroll depth and bounce rate — measured anonymously and stored only in your own site database (nothing is sent anywhere). Off by default; enable “measure engagement” to see it. A clear way to confirm the humans PureGuard lets through are genuinely engaged, while any bot that slips past shows near-zero time.
  • FIXED: The Performance view no longer counts failed or timed-out checks as “billed API calls” — they now have their own “API errors (fail-open)” figure, so your usage numbers are accurate.
  • IMPROVED: The “slowest checks” list now shows plain-English verdicts instead of internal codes.

6.0.13

  • FIXED: The Live dashboard now counts UNIQUE VISITORS, not raw page-views. A single crawler hitting thousands of pages (mostly served from cache) was inflating the “Bots” and “Checked” numbers many times over. The dashboard now shows the true number of unique visitors and bots, with total page-views shown as context — so the figures are honest and no longer alarming.

6.0.12

  • SECURITY: The connector “campaign skip” is now an explicit opt-in setting, OFF by default. On a normal website every single visitor is now verified — a visitor can no longer skip the security check by adding tracking parameters to the URL. Media buyers who send PureGuard ad-connector traffic to their site can enable “Media-buyer campaign skip” to avoid re-checking already-verified campaign traffic.

6.0.11

  • FIXED: Campaign visitors from your ad connectors are recognized across their whole visit, not just the first click. Their follow-up page views (which drop the tracking parameters) no longer trigger a redundant security re-check — removing extra latency on paid landing pages. A short-lived cookie carries the “already-verified” status; bot-looking requests never get it.

6.0.10

  • NEW: Campaign-aware Live dashboard — visitors arriving through your PureGuard ad-campaign landers (connector tracking URLs) are shown in their own “Campaign” bucket, separate from direct visitors. Their numbers now line up with your pureguard.io /live campaign stats.
  • IMPROVED: Campaign-referred visitors are no longer re-checked against the API (they were already verified upstream at the connector gate) — saves quota and removes per-visitor latency on landing pages. Bot-looking user agents never get this shortcut.
  • IMPROVED: Performance view labels now say exactly what is counted — local checks (incl. cache) vs API calls (server-billed) — so plugin totals and account totals are apples-to-apples.

6.0.9

  • NEW: “Test key” button on the settings page — instantly confirm your API key works and see your plan and monthly usage.
  • HOUSEKEEPING: Deleting the plugin now fully cleans up after itself (removes its database table and settings). Deactivating still keeps your data.

6.0.8

  • RELIABILITY: If the PureGuard service is ever unreachable, the plugin now detects it and instantly stops calling the API (no per-visitor delay) while allowing every visitor through — your site stays fast and open, never blocked. Protection resumes automatically when the service is back.
  • NEW: A clear status banner on the plugin pages tells you when protection is paused (service unreachable, or plan quota used up), so you always know the current state.

6.0.7

  • FIX: “Allow VPN” and the censorship filter now tell the engine to skip VPN detection natively, so VPN readers are reliably allowed (previously some VPNs detected as proxies were still blocked). Real datacenter / TOR / public proxies stay blocked. Verified end-to-end with a real VPN visitor.

6.0.6

  • IMPROVED: Plain-English labels in the Live dashboard (e.g. “Verified human”, “Bot user-agent”, “Proxy network”) instead of raw engine codes.
  • IMPROVED: The Campaign tab now appears only if your account actually runs traffic connectors — pure-security sites see a clean, single-scope dashboard.

6.0.5

  • PRIVACY: Engagement tracking is now OFF by default (opt-in). New installs are pure bot protection with no tracking script added; turn on Engagement in Settings → General if you want time-on-page/scroll intelligence.
  • IMPROVED: The account/campaign view now shows only your traffic connectors — your site’s own security checks no longer appear mixed in as a “source.”

6.0.4

  • NEW: Monitor mode — watch and record every visitor to your Live dashboard without ever blocking or challenging anyone. The safe first step: see your traffic for a few days, then turn on blocking once you trust it. (Off now means fully disabled — no blocking and no logging.)

6.0.3

  • SECURITY: Site protection now uses a dedicated security endpoint, fully separated from the media-buying side — your WordPress site can never be affected by ad-traffic zone reputation.
  • IMPROVED: Clearer mode descriptions — Medium (fast server-side check) vs High (adds a browser check a bot cannot fake).

6.0.2

  • NEW: Per-source breakdown in the account view — see every traffic source / connector with its own checked / accepted / blocked / accept-rate, so campaigns running side by side no longer blend into one number.
  • NEW: Source drill-down — pick a source (or click “view”) to scope the whole account dashboard to that one campaign.

6.0.1

  • NEW: PureGuard Live now has two scopes — “This site” (visitors this plugin checked at your WordPress site) and “My PureGuard account” (every visitor across all your sites and traffic connectors). The account scope shows the same numbers as your pureguard.io/live dashboard, so the two never disagree.
  • IMPROVED: Clear labels explain which checkpoint each number comes from.

6.0.0

  • NEW: Two more security modes — Strict (Humans only) and Lockdown (block everyone, admins and search engines excepted).
  • NEW: Censorship-friendly VPN filter — visitors flagged only for VPN/proxy/hosting signals are challenged instead of hard-blocked, for audiences that browse via VPN to escape censorship.
  • NEW: PureGuard Live — a full statistics dashboard with Security and Performance views: hourly charts, visitor mix, top blocked IPs/countries/reasons, live feed, API latency and cache efficiency. Powered by a local, self-pruning 30-day event log.
  • NEW: Branded gate pages — custom logo (Media Library picker), brand name, accent color, dark/light theme on both the challenge and block pages, with one-click admin previews.
  • NEW: Customizable block page — heading, message, visitor IP + incident ID + time display, and an “I am human” report button; reports appear on the Live dashboard.
  • NEW: Custom success message and optional redirect after a passed browser check.
  • CHANGED: Each site now reports to PureGuard intelligence under its own domain as the traffic source (previously a shared “WORDPRESS” source), so your site gets its own reputation.
  • FIX: Internal version constant and readme stable tag aligned.

5.0.7

  • IMPROVED: Off mode now returns before any API call — zero added latency when monitoring is all you want.

5.0.4

  • NEW: “Allow AI training crawlers” toggle (off by default). Search engines and social crawlers are always allowed.
  • CHANGED: “Allow VPN” now applies to VPNs only — open / datacenter / TOR proxies stay blocked.
  • IMPROVED: Plain-language labels (Humans / Suspicious / Bots) throughout.

5.0.3

  • IMPROVED: The plugin forwards the visitor’s browser headers (Sec-Fetch, client-hints, language) to the detection API for accurate trust scoring.

5.0.2

  • NEW: “Allow VPN / proxy visitors” toggle.

5.0.1

  • NEW: Burmese (Myanmar) challenge language. Default trust threshold aligned to 5.5.

5.0.0

  • NEW: One security mode selector (Off / Medium / High) replacing the v4 multi-dropdown setup, per-site zone identity, multilingual challenge page, stats tab, fail-open API behavior.

4.0.1

  • Compliance: challenge CSS/JS moved into enqueued assets per WordPress.org review.

4.0.0

  • BREAKING: prefixes renamed to pgperf_. JS challenge with SHA-256 proof-of-work and browser integrity checks.

3.0.0

  • Three-tier traffic classification.

2.0.0

  • WAF security layer via the PureGuard detection engine.

1.0.0

  • Initial release.