Qala Elastic: Search & Filtering, Reimagined

A faster, smarter, and more flexible search experience for your website or webshop – built for the way your customers actually search.

Status note: This document describes work that is currently scoped and in early development. The features below represent our planned direction for Qala Elastic over the coming months – they are not yet available, and timelines are still being firmed up.

Why This Matters

Search is one of the most important interactions on any website. When a visitor uses the search bar or a filter, they are telling you exactly what they want – and how quickly and accurately you respond directly impacts conversions, engagement, and trust.

Qala Elastic is our search and filtering toolkit, made up of three tightly integrated pieces:

• Qala Elastic Search – the search-as-you-type experience powering the header search bar and search results page. Built for speed, with results appearing as the visitor types.
• Qala Elastic Filters – the faceted filtering experience: checkboxes, radio buttons, range sliders, dropdowns, sorting, pagination – everything needed to help visitors narrow down what they’re looking for.
• Qala Elastic Core – the foundation tying it all together, with a single settings area for configuring how search and filtering behave across the site.

The whole suite is currently being moved from Elastic AppSearch to raw Elasticsearch. AppSearch is a friendly but heavily abstracted layer on top of Elasticsearch – it served us well, but it limits how far we can take search quality, performance, and cost efficiency. Talking directly to Elasticsearch removes those limits and is what unlocks the next generation of features described below.

---

Planned Features

These are features we have scoped and intend to build as part of, or directly following, the AppSearch → Elasticsearch migration. They are on our roadmap, not yet shipped.

A More Modern Search Experience

We plan to redesign the search results UI to feel more like the search experiences your customers already use on the world’s biggest webshops.

• Per-project visual customization – the search experience can be tailored to match the design and tone of each individual site, rather than forcing a one-size-fits-all template.
• Suggested queries as visitors type – helping them find what they want faster, and surfacing products they might not have thought to search for.
• Mixed results – products and categories shown side by side in clearly separated sections, so visitors can jump straight to a category page or a single product, whichever matches their intent.
• Larger, more inviting product images – making results scannable at a glance and visually engaging.

Search That Understands Meaning, Not Just Words (AI Search)

We are planning to add AI-powered semantic search on top of Elasticsearch. Instead of only matching keywords, the search engine would understand the intent behind a query.

• A visitor searching for “something to keep my coffee warm in the car” could find travel mugs – even if the word “mug” never appears in the product description.
• Long, natural-language queries would return relevant results instead of nothing.
• The intent is for this to work automatically: simple one or two-word searches use fast traditional search, while longer or more descriptive queries trigger the AI-powered experience. Visitors would never have to choose a mode – the system picks the right one for them.
• Configurable directly in WP Admin, so the behaviour can be adjusted per project without developer involvement.

Smart Search That Learns From Your Visitors

Search quality should improve over time as you learn more about how your visitors behave. We are planning a full feedback loop in Qala Elastic:

• Behavioural analytics – popular searches, click-through rates, what visitors click on after searching, and crucially, what they search for and find nothing (a powerful demand signal).
• An analytics dashboard in WP Admin – top queries, zero-result queries, click-through rates, and trends over time, all in one place. No third-party tools needed.

Behaviour-driven ranking – products that visitors click on and buy would automatically be promoted in future search results, with admin-configurable controls for how strongly behaviour influences ranking.

Hand-Picked Results for Your Most Important Searches (Curations)

For your highest-traffic search terms, leaving everything to the algorithm isn’t enough. We plan to build a Curations feature that gives editors complete control:

• Pin specific products to the top of results for chosen search terms – useful for promotions, seasonal campaigns, or making sure your hero product always appears first.
• Boost specific products without forcing them to the very top – a softer nudge for items you want to favour.
• All managed from a simple admin page, no developer required.

This effectively replicates AppSearch’s Curations feature, but natively on Elasticsearch and inside Qala Elastic – without depending on AppSearch.

Smarter Relevance Tuning

We are planning a redesigned relevance tuning interface that gives editors far more control over how results are ranked. AppSearch’s relevance tuning is intentionally simplified; Elasticsearch supports a much more powerful model, and the new UI is designed to expose that without overwhelming the user.

• Boost based on signals that matter to your business – popular products, best-sellers, recently updated content, etc.
• Context-aware ranking – for example, boost main products over accessories or spare parts when the query suggests the visitor is looking for the main product, without distorting results for unrelated queries.
• Tuning would happen in the admin area and changes would take effect immediately – no code deployments.

Better Reliability and Admin Tools

A number of foundational improvements are planned alongside the headline features:

• Reliable pagination, filter clearing, and OR-based filtering on Elasticsearch – polished to match (and then exceed) the current AppSearch experience.
• Immediate removal of deleted content from search results – when something is deleted in WordPress, it should disappear from search instantly. No more orphan results.
• A built-in Search Inspector inside WP Admin – administrators and editors will be able to browse exactly what is indexed, helping diagnose issues quickly. AppSearch had a friendly dashboard for this; the raw Elasticsearch / Kibana experience is much more technical, so we are bringing the convenience back natively into WP Admin.

Safe index reset tools – for cleanly re-syncing search data when needed, without taking the site offline.

---

Possible Future Features

Moving from AppSearch to raw Elasticsearch doesn’t just enable what is already on the roadmap – it opens the door to a wide range of features we could deliver in the future, for webshops, content sites, learning platforms, and any other application where search and filtering matter. These are not committed, but they all become realistic options on Elasticsearch in a way that they simply aren’t on AppSearch.

For Webshops

• Synonyms management – teach the search to understand that sofa and couch are the same thing, or that trainers and sneakers mean the same in your store.
• Personalized search results – show different results to returning visitors based on what they’ve viewed or bought before.
• Stock-aware ranking – automatically prioritize in-stock items, or push out-of-stock products further down.
• Price-range and discount-aware boosting – promote items on sale, or items in a visitor’s typical price bracket.
• Geographic and store-locator search – find the nearest store, or restrict results to products available in a specific region.
• Visual search – let visitors upload an image and find similar products.
• “People also searched for” / “Frequently bought with” suggestions – driven by real behavioural data, not static rules.

For Content Sites, LMS, and Other Applications

• Document and PDF search – let visitors search the contents of attached documents, not just titles and descriptions.
• Course and lesson search for LMS platforms – find courses by topic, learning outcome, instructor, or skill level, with rich filtering by progress, duration, or difficulty.
• Multilingual search – accurate search across multiple languages on the same site, with proper handling of word stems, accents, and language-specific quirks.
• Typo tolerance, tuned per project – generous for consumer audiences, strict for technical or legal content.
• Faceted filtering on any content type – knowledge bases, recipe sites, event listings, real-estate portals, job boards – anywhere visitors need to narrow down a large set of items.
• Date and time-range filtering – for events, news, releases, or any time-based content.
• Highlighted matches in results – show visitors exactly where in the content their query was matched.

Across the Board

• Greater scalability – Elasticsearch comfortably handles very large catalogues and content libraries.
• Better cost efficiency at scale – raw Elasticsearch is meaningfully more cost-effective than AppSearch for larger projects.
• No vendor lock-in – moving off AppSearch means we have full control over the search experience, and can build exactly what each project needs rather than working around the limits of an off-the-shelf product.

---

Why Qala

Qala is our agency’s offering – a complete suite of themes and plugins designed to work together to solve real customer problems. Every part of Qala is built with the same priorities: performance, user experience, and the needs of the people who actually use the site.

We focus on medium-sized businesses, webshops, and bespoke websites for clients across the EU – and we build everything to be both powerful for editors and effortless for visitors.

The planned Qala Elastic roadmap is the next step in delivering on that promise for one of the most important interactions on any website: helping visitors find what they’re looking for.