According to AlphaMoat, as of 2026-09, the Chrome extension PureSight has 51 users (+21.4% MoM) — ranked #5,424 of 9,324 AI browser extensions, and #69 of 123 in Other Industries.
Blocks NSFW sites and filters immodest images using AI gender detection, skin analysis, and keyword heuristics. Free + Pro. PureSight — AI Content Filter for Chrome
PureSight is the most advanced content filtering extension available for Chrome. Unlike basic blockers that only maintain a list of known adult websites, PureSight uses a triple-signal AI detection system to catch inappropriate content across the entire web — including immodest images that aren't explicitly pornographic, like swimwear, lingerie, and revealing photos on otherwise mainstream websites.
If you care about internet safety for yourself, your family, or your community, PureSight goes further than anything else available.
─────────────────────────────────────── WHY PURESIGHT IS DIFFERENT ───────────────────────────────────────
Most content filters work by blocking a fixed list of known adult domains. That approach misses the vast majority of problematic content — which appears on ordinary shopping sites, social media feeds, news outlets, Reddit, search results, and image boards that aren't "adult sites" in the traditional sense.
PureSight analyzes the actual visual content of images using three independent signals:
1. KEYWORD ANALYSIS Every image's filename, URL path, alt text, surrounding text, and parent link text is analyzed against curated lists of female, male, and safe-context keywords. This catches obvious cases instantly, before any AI model even runs — keeping performance fast even on image-heavy pages.
2. AI GENDER DETECTION PureSight uses face-api.js with the TinyFaceDetector model (the smallest and fastest available) to detect faces and classify gender with configurable confidence thresholds. When a female face is detected above the confidence threshold, the image is filtered. This catches images where keyword signals are absent — stock photos, social media profile pictures, editorial images.
3. SKIN EXPOSURE ANALYSIS A custom pixel-level skin detection algorithm (HSV + YCbCr color model) calculates the ratio of skin-toned pixels to total image pixels. Images exceeding the configurable threshold are filtered even when there's no face present at all. This is the signal that catches swimwear, lingerie, and body-focused photography that other filters completely miss.
When all three signals are considered together, PureSight achieves dramatically fewer false positives than skin-only detectors (like TAG Image Filter) and dramatically better coverage than keyword-only or domain-only filters (like FoxFilter or BlockSite).
─────────────────────────────────────── HOW FILTERING WORKS ───────────────────────────────────────
All images on a page are hidden by default via injected CSS before the page finishes loading. This eliminates the "flash of unfiltered content" problem that plagues other extensions — you never see a filtered image, even for a split second.
Each image is then analyzed through the pipeline:
→ Is this image too small to matter? (configurable minimum size) → mark safe → Is this page on the NSFW blocklist? → block the entire page → Is this page on your whitelist? → mark all images safe → Keyword analysis: strong safe signal? → mark safe. Strong female signal with no male signal? → filter → AI gender detection + skin analysis run in parallel → Final decision: filter if female face detected, or female keywords plus elevated skin ratio
Filtered images are either:
The grey overlay mode is particularly effective because it obscures the body precisely without making the whole image disappear — the page layout stays intact.
─────────────────────────────────────── WHAT PURESIGHT BLOCKS ───────────────────────────────────────
PureSight protects you from content across a much wider spectrum than other filters:
EXPLICIT CONTENT
IMMODEST BUT NOT EXPLICIT
VIDEO CONTENT
*Note on video filtering: On DRM-protected platforms like YouTube and Netflix, body segmentation isn't available due to content protection restrictions. PureSight applies blur overlay for these and uses transcript/metadata analysis. On standard HTML5 video elements across the rest of the web, full body segmentation is available.
─────────────────────────────────────── HOW PURESIGHT COMPARES ───────────────────────────────────────
BlockSite and Cold Turkey are focus/productivity tools that let you manually block specific websites. They have no visual content analysis at all — if an explicit image appears on a site that isn't on your blocked list, it passes through completely unfiltered. They are built for distraction blocking, not content safety.
FoxFilter uses keyword matching against page text and URL patterns. It has no visual analysis component, so it cannot detect images that aren't accompanied by explicit text. A fashion site product page for a bikini has no explicit words — FoxFilter won't touch it. PureSight's skin analysis and gender detection catch it anyway.
TAG Image Filter performs pure skin-color pixel detection using HSV/YCbCr color models — similar to one of PureSight's three signals. The problem with relying on skin color alone: food (certain warm-toned foods), sunsets, wood grain, sand, and sports action photos all trigger false positives. And skin detection without gender context means male athletes, fathers playing with children, and medical imagery all get flagged. PureSight's keyword and gender signals dramatically reduce false positives while catching more true positives.
uBlock Origin is a network-level adblocker. It can be configured with cosmetic rules to hide specific CSS selectors, but has no semantic image analysis. It's a powerful tool for ad blocking, not content filtering.
PureSight is the only Chrome extension that combines all three signals — keywords, AI face/gender detection, and skin analysis — in a single pipeline, with configurable thresholds, tamper protection, and a free-tier that's genuinely useful.
─────────────────────────────────────── FREE TIER — INCLUDED AT NO COST ───────────────────────────────────────
The free tier of PureSight is a real content filter, not a nag screen. It includes:
The free tier uses keyword-only image analysis.
─────────────────────────────────────── PRO TIER — $3.99/MONTH OR $29/YEAR 7-DAY FREE TRIAL — NO CHARGE UNTIL TRIAL ENDS ───────────────────────────────────────
The Pro tier adds the full AI pipeline:
─────────────────────────────────────── TAMPER PROTECTION ───────────────────────────────────────
PureSight includes a unique tamper protection system designed for users who want protection that's genuinely hard to undo in a moment of temptation.
When Challenge Lock is enabled, any attempt to loosen a protection setting — turning off the image filter, turning off the site blocker, raising the skin sensitivity threshold, disabling gender detection — requires typing a full challenge phrase before the change takes effect, without a single error.
That's 250+ characters. You can customize the phrase to something more meaningful to you.
The challenge phrase system works because it introduces friction. Turning off a filter takes less than a second normally — click, done. With the challenge, you have to deliberately type out a full paragraph. That pause is often enough.
Additionally, you can enable "Block Extensions Page" to prevent navigating to chrome://extensions — the page where extensions can be disabled entirely. This is an additional layer of protection against bypassing the filter by disabling the extension itself.
The challenge lock does NOT prevent you from tightening protection — you can always turn features on, raise thresholds to filter more, or add domains to your blocklist without any challenge.
─────────────────────────────────────── PRIVACY — EVERYTHING STAYS ON YOUR DEVICE ───────────────────────────────────────
PureSight is designed privacy-first from the ground up:
─────────────────────────────────────── SUPPORT & FEEDBACK ───────────────────────────────────────
Found an image that shouldn't be filtered (false positive)? Right-click it and select "Report: should NOT have been filtered."
Found an image that should have been filtered but wasn't (false negative)? Right-click it and select "Report: SHOULD have been filtered."
These reports help us improve the keyword lists and detection thresholds in future updates.
For support, feature requests, or bug reports, please use the Chrome Web Store support tab.
Protect your privacy and uphold Islamic values by auto detecting & blurring images and videos of unwanted or impermissible content.
+ CompareCurrency converter & calculator with live exchange rates, charts & AI price detection. 170+ currencies, private, free to use.
+ CompareMachine learning based match outcome predictor for CS2 matches on FACEIT
+ CompareL'algoritmo ti dice se conviene comprare ora o aspettare. E ti avvisa su WhatsApp e sull'email appena il prezzo scende
+ CompareData month: 2026-09 · User counts and ratings come from public Chrome Web Store data, aggregated monthly. See methodology.