Updated 2026-11-13
Blocking AI slop: eight tools already do it, and they all miss the same thing
Short answer: this one is already built, several times over. Between May and September 2026 at least eight browser extensions shipped to hide AI-generated content on YouTube and other feeds, four of them free. Every one of them relies on the same three signals: an explicit AI label, thumbnail characteristics, or channel history. All three are unreliable, and none of them can answer the question a person actually has, which is "do I trust this source." That is a different product, and it is the only version of this idea still open.
🔥 Heat 140.62
1 evidence post
1 pain point
2026-11-13
Mention 1 · 1 source🔥 Surging
At a glance
Who asks
Heavy feed consumers
People who spend hours on YouTube, Reddit and LinkedIn and notice the shift.
What they pay today
Nothing
Every competing extension is free or free with a paid tier nobody upgrades to.
First revenue
Unlikely
Extension installs convert to payment at a rate close to zero.
Cost to build v1
2 weeks
Content script plus a label heuristic. The hard part is accuracy.
Up against
8 extensions
Clarity, AiBlock, DeSlop, YTBlock, Slop Evader, AI Blocker, AI Slop Blocker, Detectly.
Ceiling
Very low
The platforms can ship this themselves tomorrow, and one of them will.
Sources: Remy's tool roundup, 2026-09-08; AI Slop Blocker launch listing, 2026-06-22; AI Content Shield listing, 2026-05-07; Chrome-Stats entry for Detectly, 2026-05-25.
What one person actually said
“I wish there was a 'block AI slob' action for the internet 😫 youtube and every Page is full of it by now... fcking annoying”
Read the shape of the wish. Not a filter, not a score, not a label. A block action, the same way an ad blocker works. One button, applied everywhere.
A single Bluesky post with 190 likes and 11 replies. That is thin evidence, and it is worth saying plainly: the demand signal here is one person's complaint plus a wave of product launches, not a measured market.
Why now
1. The word arrived, which means the perception is shared
"AI slop" moved from an insult into a category name during 2025 and 2026. It now appears in product names, in press coverage and in the names of the tools themselves. When a complaint gets a noun, it becomes a market, and the market responded within a single year.
2. The tool wave already happened
AI Content Shield shipped on 7 May 2026. Detectly was listed by late May. AI Slop Blocker shipped on 22 June as a free open-source extension that hides explicitly labeled synthetic content and runs detection locally. Clarity trains on more than 10,000 labeled AI channels. Slop'o'Meter launched in September to score LinkedIn posts from 0 to 100. That is not a gap waiting to be filled. That is a category that got crowded in five months.
Software Bolt, 2026-05-07; Chrome-Stats, 2026-05-25; prori.ai listing, 2026-06-22; Remy, 2026-09-08; Comoros Technology Digest, 2026-09-17.
3. Disclosure requirements created the only reliable signal, and it is voluntary
Labels exist because platforms started asking creators to disclose synthetic content. That is why "hide labeled AI content" works at all, and it is also why it will always be partial: a disclosure rule only catches the people willing to disclose. The tools that depend on labels inherit that ceiling, and no amount of model improvement fixes a missing label.
The gap: everyone built a detector, nobody built a trust list
The eight extensions split into three families, and each has a structural weakness.
- Label-based: AI Slop Blocker hides content that carries an explicit synthetic label. Accurate when the label exists, blind when it does not, which is most of the time.
- Detection-based: Clarity, AI Blocker and Detectly score content with models trained on channel history or thumbnails. Better coverage, and every false positive hides something a human made. The failure mode is invisible to the user, which is worse than a miss.
- Blocklist-based: AiBlock, DeSlop and YTBlock let communities maintain lists. Accurate and cheap, and dependent on someone else doing the curation, which means it lags by weeks.
- Blunt instruments: Slop Evader shows only content published before 30 November 2022. It guarantees a floor and throws away four years of legitimate work.
The question none of them answer
A person scrolling a feed is not asking "was this generated by a model." They are asking "is this worth my attention." Those are different questions, and the second one has a reliable answer that no detector can reach: the set of sources you already trust.
A product built on a personal trust list, one that follows the creators you follow across YouTube, a newsletter and a subreddit, and downranks everything outside it, solves the perception problem without pretending to solve the detection problem. It is also the only version of this idea that a solo developer can make accurate, because the user supplies the ground truth.
Reasons to build
- The complaint is felt by millions and stated by one. 190 likes under a single post is a sample of a much larger mood that shows up in comment sections everywhere. You do not need to convince anyone the problem exists.
- A trust-list product is accurate by construction. You are not guessing whether something is synthetic. You are ranking by a list the user controls, which removes the accuracy ceiling that caps all eight competitors.
- The trust layer has a natural subscription. A synced, cross-platform trust list is a service with ongoing value, unlike a one-time detector. That is the difference between a feature and a business.
Reasons not to build
- Eight products launched in five months. Shipping the ninth version of a crowded idea in a category where four competitors are free is not a business plan. The distribution math does not work: an extension that everyone can install for free has no pricing power.
- Detection accuracy is a permanent ceiling, not a milestone. Every detector will misfire. Users who lose one legitimate video to a false positive uninstall, and the reviewers write about exactly that case. You cannot ship your way past an inherent limitation.
- The platforms can do this themselves, and have started. YouTube asking creators to disclose synthetic content is the first half of a native filter. Once a platform ships "hide AI-generated," every extension in this category loses its reason to exist, and extensions cannot outrun a first-party toggle.
- Extension economics are brutal. Free installs, near-zero conversion, browser store review cycles, and a manifest change that can break your injection model overnight. The ceiling on revenue is a few thousand dollars a year, against permanent maintenance.
If you build it anyway
Build the trust list, not the detector. First version: an extension where the user imports their subscriptions and a small allowlist, everything outside is visually de-emphasized rather than hidden, and the tool never claims to know whether something was generated. Ship one platform, YouTube, because that is where the complaint is loudest and where the DOM is most stable.
The feature that makes it worth paying for is sync: the same trust list applied on a phone, in a newsletter client, and on a second browser. That is a service, and services can be charged for.
What you would charge
- Free: one browser, one platform, manual allowlist. This is what gets installed.
- $3 a month: cross-device sync and multiple platforms. Priced at impulse level because a filter is a habit, not a purchase.
- No lifetime tier. A lifetime price on a product that requires ongoing curation of trust data is a promise you cannot keep for three years.
Napkin math
Assumptions, and the conversion rate is the one that matters.
- Installs in year one from community posts and one Product Hunt launch: 20,000 (assumption, optimistic for a crowded category).
- Free to paid conversion: 1.5 percent, or 300 subscribers at $3.
- Monthly revenue: $900. Annual: $10,800, before fees and support.
Against that, you maintain a browser extension that a platform can obsolete with one product announcement. The revenue is real at this scale and the risk is structural, which is why this reads as a side project rather than a bet.
Evidence (1 post)
Bluesky, 2026-11-10. The competitor list comes from dated product listings and a September 2026 roundup, not from the post.
FAQ
Who already builds this?
At least eight extensions as of November 2026: Clarity, AiBlock, DeSlop, YTBlock, Slop Evader, AI Blocker, AI Slop Blocker and Detectly, plus Slop'o'Meter for LinkedIn posts. Four are free, one is open source.
Why do they all miss content?
Three signals, three limits. Labels only exist where a creator chose to disclose. Detection models misfire on legitimate work. Community blocklists lag behind new channels. None of them can tell you whether a source is one you trust.
What is the smallest shippable version?
An extension for one platform that downranks everything outside a user-supplied allowlist, with no claim about whether content was generated. The accuracy comes from the user's list, not from a model.
Why is a detector a bad bet?
Because the ceiling is inherent. False positives cost a legitimate creator their audience, users uninstall after one, and platforms are already moving toward native filters that make third-party extensions redundant.
Is the demand real?
The mood is real and broadly shared. The measured evidence here is one Bluesky post with 190 likes. Treat the market as unproven and the competition as proven.
How is this data collected and verified?
Competitor listings and launch dates are cited inline with source and date. The single quote links to its original post. See how data is collected.
One new evidence-backed need every day. Every trend traced to the original post and the exact quote.
Subscribe $19.99/mo →