A search box that returns the right products for the words your customers actually type.
Klevu, Algolia or Hyvä-native search fitted to Magento, Adobe Commerce, Hyvä and Shopify stores, then tuned on your real query log, with synonyms, boosts, redirects and category ordering set from evidence. For merchants whose search returns nothing for queries that should sell. Estimated in writing once we have read the log.
What merchants find when they finally read their own query log.
Someone typed the exact name of a product we sell and got no results.
The index does not know your customers' words: plurals, misspellings, trade names. Synonyms and corrections fix most of it.
Our category pages are ordered by whatever was added first.
Nobody set the order. Rules for stock, margin and popularity can be written down and changed by your team.
We pay for a search service and nobody has looked at its settings since launch.
Engines are tuned once and forgotten. Your customers keep inventing new words.
Six things, each written down so your team can see why the store behaves as it does.
Klevu, Algolia or Magento's own index, recommended after reading catalogue, traffic and budget.
Synonyms, corrections and redirects set from the queries that return nothing.
Boosts, pins and category ordering with a reason beside each, in tools your team uses.
Attribute values normalised first, so filters stop splitting one size into five.
Learned relevance configured properly, then the top queries checked by hand.
Not a chat box, and not a promised conversion figure.
A store built by this team. The figures are the store's own, measured by Google; they are not a claim about search.
A Hyvä rebuild grown into a multi-store estate for an Irish garden retailer, with ongoing development.
Core Web Vitals, real-user data (Chrome UX Report), mobile, 28 days to 31st October 2026. Marks show Google's thresholds.
Evidence first, changes on staging, then a monthly look at what people typed.
Send us the store address. A senior person tries the search with the queries your customers are likely to type and replies in writing.
Within two working daysZero-result searches, high-exit searches, misspellings and the queries that matter most by volume.
Within two working days of accessKlevu or Algolia connected and indexed against a staging copy, or Magento's own index and Hyvä search configured there.
First weekSynonyms, redirects, boosts and category rules set from the log, then the top queries tried one by one.
Second weekLive at a quiet hour agreed with you; a monthly look at new queries as part of your support, if you want it.
At a time agreed with youThree good answers. The right one depends on the catalogue, the traffic and what your own team wants to be able to change.
Magento's own OpenSearch index, served through Hyvä's search components. No monthly fee, fast on Hyvä, and better than most merchants think once synonyms, stop words and attribute weights are set from the query log. Merchandising is done in Magento's own category and rule tools.
Right for stores on Hyvä with a catalogue that is well attributed and a team happy to work in the Magento admin.A hosted search and merchandising service with its own learned relevance, natural-language handling and a merchandising console for your team. Works on Luma, Hyvä and Shopify. A monthly subscription on Klevu's published plans, billed to you.
Right for stores with a large or fast-changing catalogue, a lot of long and misspelt queries, and a team that wants to merchandise without a developer.A hosted search service built for speed and for developers, with strong rules, personalisation and analytics, and a merchandising studio. Works on Luma, Hyvä and Shopify. Priced on records and searches, billed to you directly.
Right for stores that want very fast results across several surfaces, or that already have engineers who will build on its APIs.We work with all three, so the recommendation is about your store, not about the engine.
We price the work, not the clock.
Fitting Klevu or Algolia, or tuning Magento's own search through Hyvä, is estimated in writing once we have read your query log. Klevu and Algolia charge their own subscription, billed to you directly on their published plans and not marked up by us; Magento's own search has no licence fee.
Every line of code and every piece of content that AI tooling drafts is read, tested and signed off by a senior engineer.
Years of Magento and Shopify work tell us what is safe to change on a trading store, and what needs a second look.
Whether we do it for you or your own team does it with AI, the same senior person stays answerable for the store.
It depends on the catalogue and the budget. Magento's own OpenSearch index, served through Hyvä, has no licence fee and is good enough for many stores once tuned. Klevu and Algolia are hosted services with their own monthly fee, better handling of misspellings and natural phrasing, and merchandising tools your team can use. We say which after reading your query log.
No. We can show you, in your own query log and analytics before and after, which searches returned nothing, which now return the right products and how category pages are ordered. What that does to revenue depends on your products, prices and traffic, and it is yours to measure. We do not publish uplift figures for this work.
No. Klevu and Algolia both work on Luma, on Hyvä and on Shopify. Hyvä-native search means tuning Magento's own index and serving it through Hyvä's frontend, so that option does need Hyvä.
Yes. Shopify's own search and filtering can be tuned, and Klevu and Algolia both have Shopify apps. The method is the same: read the query log, fix what returns nothing, order the rest on evidence.
We will fit one if you want one, but not first. Most of what a storefront assistant answers is a search that should have worked and a product page that should have said it. Once those are right, an assistant has something to stand on; before that, it hides the problem.
Tuning Magento's own search or re-tuning an engine you already have is usually a matter of days once we have the log. Fitting Klevu or Algolia for the first time is a short project, estimated in writing after we have looked at the catalogue and the theme.
Give us the address and we will read what the store already shows the world: release, support status, PHP, headers, hosting and any sign of a hijack. One page, by email, within two working days. Nothing is accessed or tested.
What we would do first. Apply the outstanding security fixes, then plan the upgrade to a supported release.