🚨 EXPOSING NOSTRA. AI 🚨 We have exposed some pretty nasty grifts in this space, but Nostra reigns supreme above all others (by a fair margin). Since we have >30 min video and an intensive Notion document (linked at the very bottom of this post) I am going to just highlight the key areas below. For those who don't want to watch it all - here's some time stamps that cover the most important/most hilarious parts of the video. 1:08 - Site Speed Scamming 101 4:20 - Beginning of the actual findings of what we caught Nostra doing. 5:25 - Nostra CEO tweets about how vitally important it is to have your most critical information above the fold on your site. Then we show that they almost exclusively marketed their fake site speed scores above the fold. 6:37 - The 'before and after' that shows how radically different the Nostra team made their site after they realized they were being investigated (perhaps my favorite part) 11:05 - Before: YOUR SITE PERFORMANCE SCORE MATTERS ..... 1 day later, Nostra CEO: "Yeah page speed scores are pretty useless" 😂 12:07 - "Nostra clients are 8x more likely to pass core web vitals compared to non Nostra clients" *Lukas then shows how every single one of their 'success story' case studies are failing almost all core web vitals on their home pages.... 15:10 - Jake dives into the technical shortcomings of the Nostra product, from them using a deprecated form of rendering and claiming (falsely) that Google still endorses it, to showing that any visitor logged into any Nostra-enabled site is not fed the cached pages... meaning that many of their highest LTV clients are being given a drastically worse browsing experience.... lots more! 29:47 - Lukas shows all of the deleted tweets from the Nostra CEO. Suspiciously, all of them just so happen to be based on page speed scores... hmm. The main points: - Nostra uses site speed cloaking tactics to artificially inflate performance numbers on Google's Page Speed Insights. There is no ethical reason to do this. It is a tactic used exclusively by site speed scammers. *A few people may point to the fact that Meta uses Lighthouse scores as one of many contributing factors to showing ads, so certain sites MAY see an uptick in paid performance on Meta while having these fake scores. This is a horrible basis to justify attempting black hat scams on Google tools. I'll be doing a whole separate video for this topic alone, but for now, just understand it is shortsighted and unbelievably stupid. - They then used those artificially inflated performance scores as the central focus of their entire marketing strategy. *They have released some hilariously inaccurate/misleading blog posts in the past week that try and claim they don't use cloaking and that their methods are totally ethical... you better believe we are doing a follow-up video that dismantles these blog posts paragraph by paragraph. - Various current and former clients of Nostra have confirmed that one of their central selling points when convincing them to pay for Nostra (often quoting/charging thousands a month) was that their Google performance numbers were going to go up, which meant a faster site, which meant more revenue. A blatant, irrefutable lie. - As we dug into their code, we found even more issues. Most notably, their 'crawler optimization' was stripping down pages for both Lighthouse (Page Speed Insights) as well as Googlebot, which means that the contents of any page 'client-side rendered' by Nostra in this way was almost entirely invisible to Google, as it saw basically nothing to crawl and index. - All of our findings were confirmed by over a dozen well-respected developers in the Shopify space, including high-ranking engineers at Shopify. - Before we notified Nostra of our investigation into them, Jake Casto (partner in this report) met directly with the Nostra team, including their Chief Lead Architect (?) and asked every question he could to gain as much context as possible. This meeting further confirmed all of our findings and even pushed some further. - Hours after we notified the CEO of Nostra about our investigation into them and the impending report we would release, their entire site changed.... like... CHANGED. Nearly every mention of 'page speed' or 'performance score' was stripped from the site, including all of their case studies. Additionally, they renamed an entire product. Their 'Crawler Optimization' tool became 'Bimodal Dynamic Rendering'.... - That same night, the Nostra CEO then deleted all tweets insinuating performance score as a benefit of using Nostra (proof shown in the video) and began publicly talking about how useless speed scores are. A metric they had long lauded as the single-most important aspect of what their tech improves was now "pretty useless" just a few hours later. - It is imperative to know that we did not mention ANYTHING about our interest in investigating their focus on performance scores as a key marketing strategy. All of these changes were made by them without knowing anything about what in particular we were investigating. Not shockingly, the main scam we were highlighting in our investigation is what was wiped entirely (within literal hours) from their site/their founders personal messaging. * go look at their site now and try to find any claims about performance scores on their home page. They even took them off the top of all their case studies. (we show this all in the video as well). - Nostra has since continued to modify their code, resulting in some of their 'success story' clients seeing a 50 point drop in performance scores (shown in the video). More current and former clients continue to reach out and share more stories about the many sketchy happenings at Nostra. - Nostra also released a very weak response in the form of 2 blog posts that aim to justify their actions/tech. They have been sharing this with their clients and attempting to patch over the MANY inconsistencies and blatant lies they were caught in. As I said earlier, we will be doing a video dedicated to dismantling these blog posts in detail. Moral of the story. No SaaS is going to plug in to your Shopify store and drastically increase your performance scores in a matter of moments. Any tool or dev or agency, no matter how fancy they look and how much venture backing they have, will be able to get your Shopify stores' mobile performance scores into the 80/90's under any normal circumstances. If someone says they can... You are 100% being scammed. Site speed optimization is a complex development process that takes highly-skilled devs dozens of hours to do properly. No tool can replace this. Don't be fooled into thinking otherwise. For a deeper look into the technical side, check out the link below. nostra-deep-dive.com

Big rollouts this month: - Brand new analytics and home page, over 100 new metrics and dimensions. LayersQL is back up to 1:1 of ShopifyQL - Bulk editing everywhere. - Managed Collections 🤫 - Functions (using Shopify runtime/sandbox) for product data manipulation. Replaces our JSONLogic. - An Audit log of what our AI models do in search. - AI driven color filtering. - Agent Feedback API 🤫 - Brand new filtering and facets experience.

CLAP FOR US

An EVP of one of the biggest brands on @Shopify soaking highly of @uselayers_ is a feeling I will never grow old of. Legacy search & merchandising tech is dead. DMs open :)

I don’t think ecommerce websites are disappearing. But I do think they will no longer be the first step in the shopping process. There is a lot of excitement around new shopping form factors, whether that means conversational interfaces, voice assistants, wearable devices, or agents that buy things on your behalf. All of those will likely exist. But the structural change happening right now is simpler and more immediate. 1. The “shopping around” phase is disappearing A few years ago, a typical ecommerce behavior looked like this: someone searched a product name, opened five tabs, compared the same SKU across different sites, and then picked the best option. That behavior is becoming rare. When we analyze referral data across many stores, it is very uncommon to see someone arrive from another brand’s website. They arrive from search engines, from ads, or from recommendation systems. That suggests the comparison step is already happening somewhere else. 2. Search engines and AI assistants will do the comparison Traditional ecommerce browsing assumed that the customer would perform the comparison themselves. They would visit multiple sites, check prices, evaluate options, and then decide. Now those comparisons are increasingly being handled by algorithms that present the top few choices immediately. 3. The brand site becomes the destination, not the discovery layer Even if discovery happens somewhere else, the brand website still serves a unique purpose. Customers who care about a specific brand tend to go directly to that brand’s site because that environment communicates trust, identity, and product depth in ways intermediaries cannot replicate. It would not make sense for a search engine or an AI assistant to generate a full brand experience for every company. AI may compress discovery and comparison into a single interaction, but the brand relationship still lives on the brand’s own property. The future probably looks like this: AI helps customers decide what to buy. Brand websites remain where they decide who to buy from.

👀👀👀👀👀

You asked so I wrote! How we ensure 100% integration test coverage of our Shopify API integration medium.com/engineering-layer…

The gap between us and all legacy search and merch tools is widening at a crazy pace…

Probably the best feeling you’ll have and it comes at different times in your life But you need to be reminded that no one really gives a shit what you do as long as you Don’t steal Don’t lie Don’t speak on shit you don’t know

Shoutout legend @yojimmykim for the @uselayers_ plug 🙏 If you're a Shopify Plus brand using Algolia, Nosto, Boost, Athos/Searchspring, Fast Simon or Bloomreach... you're paying way too much for a vastly inferior search & merchandising platform. Let's talk.

hello I'm the non-technical co-founder and I am absolutely losing my shirt at how cool this is...

🛟🛟🛟 uselayers. com

Dinosaurs will die. Every. Single. Time. I have gone against Algolia in the past few months, Layers has absolutely smoked them, at > half the price, mind you. Without fail, I end up leaving the call equal parts shocked, confused, and excited. How can it be THAT bad? It can't possibly be? But... I just saw it with my own eyes and heard it from an actual user of the platform... We're going to absolutely run the field against these geriatric search & merch tools. This is what happens when massive barges of a company try to steer into extremely fast and dynamic markets like ecom, let alone @Shopify. It's not even fair. Let it be known. If you're using Algolia (or Nosto/Fast Simon/Bloomreach etc) I would LOVE to go head-to-head against them and show you what an enterprise-grade AI search and merch platform can really be in this the year of Our Lord, 2026. DMs are open.

stop. read this. understand it. your whole world can change in 5 years. 2026 is a week away. by 2030, you can have everything you want. yesterday, I sat in the backyard of a 25 million dollar house. I ran a panel with 3 billion in annual sales on it. the audience was 100 founders, with 20 billion in annual sales between them. most of them bootstrapped. The person who owned the house? dead broke 5 years ago. 5 years ago, connor and I shared a two bedroom apartment in venice. his bike got stolen. our girlfriends moved in. his fucking cats kept fucking up my bathroom. 5 years later- ridge is bootstrapped, growing faster then ever, and we control our fate. 5 years can change everything. you deserve to change everything.

Some of the best decisions we made at @uselayers_ were around what we didn’t build. 1) The first principle we agreed on early: performance should be our first competitive advantage. When you enter a crowded market, there’s a temptation to differentiate through features. We decided to differentiate through speed. Even if it’s just 50 milliseconds faster, that’s still a story you can tell. And when traffic spikes and a hundred thousand shoppers hit your store at once, that 50 milliseconds compounds. Your pages still load, your search still returns, and your checkout still flows. 2) The second principle: stay fast, even when it hurts. For months, we knew our search wasn’t as feature-complete as others. We didn’t have all the advanced modeling or post-processing pipelines that made other engines look “smarter.” But we refused to bolt them on because doing so would’ve made our API take two seconds to respond. We were okay being “less competitive” for a while if it meant staying fast. We even turned away customers. Literally told them to go to other platforms if they needed heavier workflows. And that patience paid off. Today, we’re running real-time query expansion and intent processing, with the same speed advantage we had from day one. When I look at some of the “RAG-based” Shopify search products today, they proudly say their retrieval workflow completes in under three seconds. Three seconds. No one’s waiting that long for search results. Our pipeline runs similar logic in half a second. Real-time, in full context, without the weight of complexity. By resisting the urge to complicate, we built something that stayed fast and became smarter over time. That’s the paradox of building products: every shortcut to “more advanced” is usually a long road to slow.

Is it cringe to say uselayers. com is all of the above? 🙈

DO IT

I am happy to announce that I am making moves in silence. Pleased to announce that motions are being made stealthily. Excited to say that things are occurring in quiet.

Lots of initial takers but we need more!

tf is Padel?

May I introduce the ‘not interested’ button to you, sir. Very effective.