Queryra
AI-powered semantic search API
WordPress.org rotates eight lesser-known plugins into the tab that opens by
default when you go to install anything. No application, no fee. Someone there
picks eight every two weeks. Mine got picked.
Two weeks of the best free distribution I am ever going to get.
I recorded a baseline before it started and logged the numbers daily, because
I could not find a single founder who had been through this and published what
it actually did.
Installs went from 30 to 200.
Then the part I did not expect. The share of those people who ever used the
product went from 50% to 4.8%. In absolute terms: thirteen people used it in
the thirteen days before, twenty-seven during.
Twenty-two times the traffic bought twice the users.
I keep coming back to that one. Not because it is bad news, but because for two
weeks I would have called this a success and been right by every number I would
normally have quoted.
The people were real. Email verification held at the same rate on either side,
so it was not scripts. What changed was why they arrived. Before, somebody
installing had gone looking for what I make. During, they were scanning a list
of eight on a screen WordPress opens by itself.
Discovery and intent are not the same thing, and only one of them shows up in
an install count.
The thing I would tell anyone about to get a spike: decide before it starts
which number would make you call it a success, and make sure that number sits
downstream of the install. Otherwise you get the spike, feel great, and learn
nothing.
All of it is written up, including the half that does not flatter me:
https://queryra.com/blog/wordpress-org-featured-plugin
Has anyone here had a distribution spike and measured what came out the other
end? I have one data point and it is my own, which is the least useful kind.
I build an AI search plugin for WordPress. One person, no funding, small.
Last week WordPress.org put it in the Featured tab. That is the tab that opens by default when you click Plugins → Add New, on every WordPress install. Eight plugins at a time, rotating every two weeks.
The part that surprised me: there is no application form and no fee. Every eligible plugin sits in the pool automatically and someone at WordPress.org picks eight. You cannot buy your way in. I found out the same way everyone else does, by looking.
Then I went looking for prior art. The programme has been running since March, dozens of plugins have been through it, and I could not find a single author who had published what it did to their numbers. The curator posted figures for the first cohort in March and said more would follow. There was no second report. Someone asked publicly for a list of previously featured plugins and the question sat there unanswered.
So there is a distribution channel that a lot of small WordPress developers wonder about, and no honest first-hand account of what it is worth.
I recorded my baseline before the rotation started. Every day I am logging downloads, active installs, signups, and how many of those people ever run a search. That last number is the one I actually care about, because installs are easy and usage is not.
One thing I have already found that I do not think is documented anywhere: the order of the eight plugins in the tab is reshuffled every day. Same eight, different sequence. So position in that list is a draw, not a ranking. Which incidentally hands me a free randomised experiment on whether position matters at all.
I will publish the whole thing when the fortnight ends. Including if the answer is "not much", which is a real possibility.
Different search engines, different catalogues, one result in common.
Try this on your own site search:
"X without Y" — X is something you sell, Y is something you don't want
then just "X"
On four of the five sites, the first query returned MORE results than the
second. Not fewer. On three of them, the thing the customer excluded was
promoted to the top of the page.
One site went from 1,340 results to 4,607. The top four hits were all the
excluded term.
Nobody is hiding this. One vendor documents it plainly: if you don't supply
an operator, the default is OR. So "lobster rolls but no Maine" becomes
lobster OR rolls OR but OR no OR Maine. The word your customer used to
exclude something quietly becomes another word to search for.
Which is fine, right up until a customer types a sentence instead of a noun.
Fair disclosure: I build a search plugin, and mine doesn't solve this with a
filter either. It rewrites "without X" into a positive form, which works when
the catalogue has something positive to match and drifts when it doesn't.
Better, not solved.
Run the two queries on your own site. Thirty seconds and you'll know which
kind you have.
Eight months of solo building. This week my AI search plugin became the #1 result for "semantic search" in the official WordPress plugin directory — ahead of plugins with install bases hundreds of times larger.
It's an algorithmic ranking, not an award, and it wobbles daily. But what got it there wasn't marketing: answering every support thread, shipping small releases, and a readme that's honest about limitations. Directories seem to reward plugins that behave well — which is encouraging if you're small.
Happy to share details about building in the WordPress ecosystem solo.
-
2
The interesting part is that the ranking seems to be a consequence of consistently serving the existing users rather than an SEO tactic. It’s a useful reminder that in a directory-driven ecosystem, support and product quality can compound into distribution.
-
1
Exactly how I see it. I didn't plan this as an SEO strategy — most of the time I was simply focused on making the product work well and making sure anyone who installed it could get a quick answer if they needed help.
Only later did I notice how those small things started to compound: better product → better user experience → more activity → more visibility in the directory.
And I think that's why the #1 was more interesting to me than the install count itself. It shows that in a directory-driven ecosystem, you can sometimes punch above your weight through product fit and consistent execution, even without a large user base.
-
2
That’s a strong example of compounding from product quality. Has the #1 ranking led to any noticeable change in installs or inbound interest yet?
-
1
I just checked — it's already back to #3 😅 The ranking is very dynamic, so I don't expect a jump that you can really notice from one day to the next.
In my first month I had just one user. Now I'm already close to 10 new installs per week — and once I hit that, I'll definitely celebrate. 😄
The logs show that there's more traffic and interest, but WordPress.org doesn't give me anything like Google Analytics where I can see how often a specific phrase was searched or how many people saw my listing.
So I'm sure it is working, but the effect isn't obvious enough to see from one day to the next.
-
1
Nice, and congrats on getting close to 10 installs/week. If you’re open to it, what’s the best email to reach you at? Would be good to stay in touch.
-
-
-
-
I'd been marketing the wrong half
I build an AI search engine for WooCommerce stores. Over four months
my API logged 155,730 queries with actual text in them.
Under 3% came from a real store. The rest was bots hitting public
search endpoints. One source alone sent 65,000 queries.
But the 3% is what got me.
Every demo I have ever shown uses a sentence. "Something for oily skin
that isn't greasy." Those queries are real, I have them in the log.
Then I checked WHEN they arrive: clustered in the first day or two
after a store connects, in a burst, from one location.
That is the shop owner, testing. Not a customer.
What real visitors type, weeks later, steadily:
drifter salts
xros 5
epson emc800
Names, model numbers, part codes, usually misspelled. I built the
sentence half beautifully and spent months marketing it, and it is
the smaller half of what my customers actually get.
The other thing is worse. I found a query that came back empty on
eight separate attempts and concluded the store did not sell that
product. It did. I had not finished importing their catalogue. I was
one step from writing that up as a customer insight when it was a gap
in my own pipeline.
Same store, three different zero-result rates depending on what I
stripped out: 48.9%, then 15.3%, then under 7%. Only the last one was
about customers.
Two things I would tell anyone building on top of user data.
Your logs are full of machines, and machines are indistinguishable
from insight until you filter them out.
And check your own pipeline before you conclude anything about your
customer. The most confident wrong answer I produced this year came
from data I trusted because it was mine.
Full write-up:
Everyone's talking about Google AI Mode as an SEO problem: will my products still show up in the answers?
That's real. But I think the bigger, quieter shift is happening somewhere else, on a surface store owners actually control.
AI Mode answers full questions. You type "a warm waterproof jacket for hiking under £150" and Google just answers it. After doing that a few times, people stop stripping their searches down to two keywords. They start typing full, natural sentences.
Here's the part I keep coming back to: that habit doesn't stay on Google. It follows shoppers straight into the search box on your store.
So they arrive and type "a moisturizer that isn't greasy" or "a lightweight day cream under £25." And default WooCommerce search, along with most plugins wearing an "AI search" label, just match keywords. The negation is invisible. The budget is treated as two words. The shopper gets "no results," decides you don't stock it, and leaves. Even when you had exactly what they wanted.
I build semantic search for WooCommerce (my project is Queryra), so I see this gap constantly. The interesting thing isn't the tech, it's that the bar just moved without anyone announcing it. Shoppers now expect a search box to understand a sentence, because Google trained them to.
I wrote up the full argument plus a one-minute test you can run on your own store, whoever's search you use. Type a full sentence with a constraint or a negation into your search box. If you get nothing useful, your search is keyword-based, and it's about to feel broken.
There's also a public demo you can poke at without signing up, if you want to see what the other category actually does (it even handles "krem do twarzy" on an English store).
queryra.com/blog/google-ai-mode-woocommerce-search
Curious whether other founders are seeing the same shift in their own search analytics.
For the last 6 months I've been building Queryra, a
semantic search plugin for WooCommerce. Solo, bootstrapped,
research phase. No paying customers by design (the "100
free partners vs 1 customer" model).
Last weekend I did something I'd been putting off:
I systematically tested every "AI search" competitor in
the WordPress directory. Wanted to verify what they
actually do under the hood vs what their marketing claims.
What I found was clearer than I expected.
FiboSearch. Keyword matching with fuzzy + synonyms.
The "AI" comes from an optional ChatGPT integration on top.
SearchWP. Keyword-based with engine weighting. AI
features (semantic, image, suggest) are paid add-ons
calling OpenAI for query refinement. The core is keyword.
Relevanssi. Full-text search with custom weighting.
Doesn't claim AI. Refreshingly honest.
AI Vector Search (Semantic). The most interesting case.
Default Lite Mode uses TF-IDF (keyword). Real vectors only
with paid Self-Hosted Supabase + your own OpenAI API key.
Queryra (mine). Real vector embeddings via ChromaDB
plus LLM intent parsing. Free tier. 50+ languages tested.
The marketing-vs-reality gap is wider than I assumed. Most
"AI search" plugins are keyword search with a marketing
layer on top.
This affected how I think about positioning. I'd been
playing nice. Calling Queryra "another option in the AI
search space." Now I realize: the space isn't crowded with
AI search. It's crowded with keyword search wearing AI
labels.
That's a different competitive landscape than the one in
plugin marketing copy.
For solo founders building in mature categories: it's
worth verifying that the "competitors" in your space
actually do what they claim. Sometimes the apparent crowd
thins out when you check the technical reality.
Wrote up the full methodology + side-by-side tests for
anyone evaluating AI search plugins:
https://queryra.com/blog/which-wordpress-ai-search-plugins-actually-use-semantic-search
Curious if other IH folks have found similar gaps in your
categories. Where do marketing claims and product reality
diverge most in WordPress / SaaS space?
-
2
One thing stood out to me.
Your research seems to answer why Queryra is technically different, but I'm not sure it answers why the first customer buys.
Those are often different decisions.
The risk is spending months proving the category is misrepresented while the real bottleneck sits somewhere else entirely.
I wouldn't try to solve that casually in the thread because the answer affects positioning, validation, and what signal actually matters before revenue.
If you're open to it, share your email and I'll put the tighter read together properly.
-
1
Fair observation — those are different decisions.
For context: Queryra is in research phase by design, no
paying customers yet (the "100 free partners vs 1 customer"
model I mentioned in the post). The blog post addresses
positioning vs the category, which is upstream of customer
acquisition.
What signal would you want to see before revenue? Genuinely
curious — happy to discuss in the thread.
-
1
Fair. That’s exactly where I’d be careful.
Research-stage products can easily measure interest in the thesis, while the real test is whether the buyer has enough pain to change behavior.
The right signal depends on who Queryra needs to become indispensable for.
I wouldn’t answer that loosely in-thread because the wrong signal can create false confidence.
If open, send me your email and I’ll write the tighter version properly.
-
-
1. Google is gradually shifting search from keyword matching to AI-driven answers. AI Overviews baseline now, AI Mode expanding this summer. Old "10 blue links" becoming the exception.
2. User behavior backs it: longer dwell time on AI-driven results, lower bounce rate, even vague queries returning relevant outcomes. Modern shoppers describe problems ("something for back pain at the office"), not SKUs.
3. Same shift visible on WooCommerce / Shopify stores adopting semantic search: visitors finding products with vague queries, conversions improving, cart abandonment dropping on search-driven sessions.
Old keyword search worked when users knew exactly what they wanted. That assumption is breaking down.
If you're building anything in e-commerce, AI search, or product discovery, this trend is worth watching. Semantic search is no longer experimental, it's becoming baseline expectation.
What signals are you seeing in your niche?
Solo dev building Queryra (AI semantic search plugin for WordPress/WooCommerce). Just got listed in the official integration docs of Oxygen AND Breakdance — two of the biggest WordPress page builders.
The distribution problem:
I had 10+ active installs on WordPress.org but couldn't break through the visibility wall. Reviews require installs, installs require visibility — classic chicken-and-egg. So I started looking for permanent placements outside WordPress.org.
The strategy:
Cold email to Oxygen and Breakdance teams. Both are owned by Soflyy, so technically two emails to the same company. My pitch wasn't "please list us" — it was "here's a complete integration doc I wrote, formatted for your CMS. You just need to review and publish."
What worked:
— Personal intro (who I am, what Queryra does)
— Confirmed compatibility status (empirically tested on their specific versions: Oxygen 6.0.0 stable + Breakdance latest)
— A finished doc draft they could paste into their editor
— Clear CTA: "Should I prepare the same for [other builder]?"
The whole exchange took ~2 weeks. Both docs are now live.
Links:
— Oxygen integration: https://oxygenbuilder.com/documentation/integrations/queryra-ai-search/
— Breakdance integration: https://breakdance.com/documentation/integrations/queryra-ai-search/
— My own integration docs (more detail): https://queryra.com/docs
— The deeper story on why page builder content matters for search: https://queryra.com/blog/page-builders-ai-search-wordpress
— Plugin: https://wordpress.org/plugins/queryra-ai-search
Lessons:
1. Remove ALL friction from the other side. Don't ask them to write content — write it for them. Format it for their CMS. Make publishing a 5-minute task on their end.
2. Page builder communities are underutilized distribution. Every major builder has an integration docs section. Most plugin developers never approach them.
3. Permanent backlinks > one-time blog mentions. These docs will rank in Google + AI engines for years. Already seeing inbound traffic.
Bonus signal: a few days after the docs went live, ChatGPT and Perplexity started surfacing Queryra more frequently when people ask about "WordPress AI search" — measurable AEO lift from authoritative domains linking to me.
Anyone else found unconventional distribution channels for plugins/SaaS? Curious what's worked outside the obvious channels (Product Hunt, Reddit, paid ads).
Before I paid, I decided to check whether these reviews actually work. Spent three weeks pulling Wayback Machine snapshots for 24 plugins across WP Mayor, Barn2, ThemeIsle, WPExplorer, and LearnWoo — install counts before and after each review.
Results: 15 out of 21 plugins showed zero measurable install growth.
3 plugins actually declined. The only site where any plugins showed organic growth was ThemeIsle — which explicitly refuses paid reviews.
A few things that surprised me along the way:
WP Mayor CEO Mark Zahra responded before I published and made genuinely fair points. His argument: their clients sell premium plugins, conversions happen through direct sales — not WordPress.org free installs. I included his full response.
Barn2's own transparency reports showed their sponsored content revenue dropped 82% in one year. They quietly removed the advertising page.
The WordPress.org ecosystem is basically a closed loop — most users discover plugins through the admin dashboard search, not blog articles.
I'm building Queryra (AI semantic search for WordPress/WooCommerce) and this research saved me probably $1,000+ in marketing spend I would have wasted.
Full data and methodology here: https://queryra.com/blog/paid-plugin-reviews-roi-research
Has anyone here had different experience with paid editorial placements?
2 Comments
Nothing clean yet — email verification held at exactly the same rate on both sides, so that one's out. Best candidate is whether they sync a catalogue, since it's the first step that costs them anything.