Product recommendation engines choose which items to show on product pages, carts and emails, based on what other shoppers did.
This guide ranks fourteen vendors on set-up effort, what the invoice becomes once modules and usage meters add up, how easily you can leave, and whether a small independent vendor can serve you without a group behind it.
Vendors can pay for more visibility on this website, for example a higher place in a ranking or more room on a page. Commercial relationships can also affect which products we feature and how prominently. Every vendor link is marked the same, whatever the arrangement behind it. How that works.
In short
What product recommendation engines software does
A product recommendation engine turns shopper clicks, purchases and catalogue data into suggestions such as similar items, bundles and trending products, shown on a shop's pages, carts and emails.
Ranking criteria for this product recommendation engines tool list, applied top to bottom: effort to get started, cost in practice, data return, ability to leave, buyer fit. Vendors outside the best-known names fill places 1 to 3 when they exist, and the big brands rank from place 4. Why those five, and why there is no score out of ten, is on the how we work page.
Request pricing; custom quote scaled by audience, products and volume
—
Large retailers wanting recommendations bundled with search and marketing automation
Small stores that need monthly billing or a trial first
Country is where the vendor is headquartered or contracts from, which is a different question from where your data is hosted. Where the two tell different stories, the entry says so.
Usage-metered recommendations from a Prague company, with a permanent entry tier
Ranked #1 of 14 in Best Product Recommendation Engines in 2026.
Free tierPublished pricingElsewhere
Plan limits are counted per month across detail views, purchases, ratings, cart additions, bookmarks and view portions, plus requests, active users and catalogue items, in tiers from under 100,000 up to 20 million or more.
Terms are Czech-law and signed with Recombee, s.r.o. in Prague. That makes it the cheapest way to prove value, but the invoice tracks traffic, so seasonal peaks move you up a tier.
What stands out
Usage tiers
Developer-led
Prague company
Where it costs you
The pages read describe an API reference and a query language, so a team without a developer will struggle to use it.
Four separate meters (interactions, requests, users, items) make the monthly cost harder to predict than a single order count.
Right for
Teams with a developer wanting published tiers and a fallback plan
Wrong for
Shop owners wanting a plugin installed without any engineering work
Czech RepublicPublished monthly tiers by usage; free plan after a 30-day trial; top tier quoted
Dutch merchandising and recommendations tool with a Basic tier
Ranked #2 of 14 in Best Product Recommendation Engines in 2026.
Free tierPublished pricingEurope
The pricing page lists Basic, Advanced (10 A/B tests a year, unlimited Boost & Bury rules, search banners, visual components, direct Google Analytics 4 connection) and Expert, which adds AI personalisation.
Paid tiers show a Get pricing button and the amounts carry no currency symbol, so confirm the currency before comparing. Recommendations are one solution among several, so check how much of the tier you would use.
What stands out
Merchandising-led
Three tiers
Zwolle company
Where it costs you
Basic caps A/B tests and Boost & Bury rules at three each per year, so active testers outgrow it quickly.
AI personalisation is listed only on the Expert tier, the most expensive step.
Right for
Merchandisers who want a tiered price ladder and a no-charge starting point
Wrong for
Developers wanting raw recommendation endpoints rather than a merchandising interface
NetherlandsFree tier called Basic, then published from-prices for Advanced and Expert tiers
Search, recommender and listings in one Bratislava product with two plans
Ranked #3 of 14 in Best Product Recommendation Engines in 2026.
Pricing on requestEurope
Integrations listed are Shopify, WooCommerce, Magento, Shopware, PrestaShop, OpenCart and BigCommerce. The pricing page says every plan carries the full feature set, so the choice is Growth with guided integration and premium support or Enterprise for large, complex catalogues.
Pricing follows usage rather than domain count. The pages read do not show the Recommender priced alone, which matters if you already run a separate search engine.
What stands out
Recommender module
Full feature set
Seven platforms
Where it costs you
Both plans show Contact Us instead of a price, so budgeting waits for a sales call.
The trial button states no length, so you cannot plan a test window in advance.
Right for
Shops on one of seven named platforms wanting search and recommendations from one vendor
Wrong for
Buyers who need a published price before they speak to anyone
SlovakiaQuoted per plan on usage, extra domains included; trial button on the site
Danish recommendation and search products priced separately by usage
Ranked #4 of 14 in Best Product Recommendation Engines in 2026.
Pricing on requestEurope
Usage tiers count monthly searches, recommendations, conversations, opened emails or active subscribers, and the page offers GBP, DKK, EUR and USD billing. A separate B2B path promises customised platform set-ups with dedicated support.
The terms follow Danish law with the Copenhagen city court as venue, while the copyright line names a different Inc., so check which entity signs your contract.
What stands out
Per-product pricing
Copenhagen ApS
B2B path
Where it costs you
Each module shows Contact Us for Quote, so a multi-product bill cannot be added up from the page.
The trial button gives no length, which makes a proper test window hard to plan.
Right for
Shops adding recommendations first and email or search later, paying per usage
Wrong for
Anyone needing a fixed monthly figure before talking to a salesperson
DenmarkUsage-based per product with quotes on request; trial button on the site
Onsite recommendations and cart-recovery tools on rolling monthly terms
Ranked #5 of 14 in Best Product Recommendation Engines in 2026.
Published pricingMiddle East
The page lists a starting price covering onsite personalisation and recommendation widgets, with other plans sized on average monthly traffic or annual sales and a Contact us route for a personalised offer.
Plans can be switched at any time. Every plan carries a customisable interface and a full analytics dashboard. Rolling terms lower the cost of leaving, but the entry price is paid every month while you are deciding.
What stands out
Month to month
Human support
Israeli company
Where it costs you
Testing means paying the entry price for a month, because no trial appears.
Plans beyond the entry tier are tailored by traffic or annual sales and need a conversation.
Right for
Mid-sized shops wanting monthly terms with a named support person
Wrong for
Teams that expect to try the product before paying anything
IsraelPublished entry price for personalisation, other plans by traffic or sales; monthly, no long contracts
Shopify recommendation app with one flat plan and a 15-day trial
Ranked #6 of 14 in Best Product Recommendation Engines in 2026.
Published pricingElsewhere
The home page promises AI product recommendations, upsells and bundles across website, mobile and email, and says no developer is needed. The pricing page lists one plan, Max, with all features and unlimited LimeSpot revenue uplift, and states a lifetime discount for signing up during the offer.
Customisation at installation is included. The terms name LimeSpot Solutions Inc. and its affiliates but no country, so ask where the contract is governed.
What stands out
Flat plan
Shopify app
15-day trial
Where it costs you
A revenue ceiling on the listed plan means a growing store outruns the published price.
The terms give no address or country, so the contracting jurisdiction is unclear.
Right for
Small Shopify stores wanting one predictable monthly price and hands-on set-up
Wrong for
Large catalogues or retailers selling outside a Shopify storefront
Not stated by the vendorPublished flat monthly plan with all features; 15-day trial
Shopify personalisation modules priced by order volume, bought one at a time
Ranked #7 of 14 in Best Product Recommendation Engines in 2026.
Published pricingElsewhere
The about page says two developer brothers founded Rebuy in 2017. The calculator asks about average monthly Shopify orders, then adds modules, each with its own price.
Buttons offer a trial and a demo, with no trial length given. Monetize is opt-in and pays monthly. Treat the as-low-as figure as the lowest module, not a typical bill.
What stands out
Order-based bands
Modular
Shopify only
Where it costs you
Every page read targets Shopify, so a store on another platform finds no route in.
Four paid modules are priced separately, so a full set costs more than the starting figure shown.
Right for
Growing Shopify brands whose cost should follow order volume, module by module
Wrong for
Retailers on Magento, WooCommerce or any non-Shopify platform
Not stated by the vendorPublished, banded by monthly Shopify orders, modules priced separately; trial button
Recommendations, bundles and search sold in three plans with a 14-day trial
Ranked #8 of 14 in Best Product Recommendation Engines in 2026.
Pricing on requestElsewhere
The site lists recommendation, dynamic bundles, search and discovery, and cross-channel marketing as solutions. The trial is promised as live within five days. Details of the three plans come from the sales manager.
Async JavaScript is described as preventing slowdown of the website, which is a vendor claim rather than a figure. Because no country is stated, check the contracting entity in the order form.
What stands out
Three product lines
14-day trial
Async script
Where it costs you
Prices appear nowhere on the pages read, so a comparison needs a sales contact.
The site names no country, so legal jurisdiction has to be asked before signing.
Right for
Retailers wanting recommendations with a short trial before any sales call
Wrong for
Buyers who need list prices and a stated company location
Not stated by the vendorQuoted by product line through a sales manager; 14-day trial
Shopify search app that adds recommendations, bundles and back-in-stock notices
Ranked #9 of 14 in Best Product Recommendation Engines in 2026.
Pricing on requestElsewhere
The feature list covers predictive bundles, pre-order, a back-in-stock Notify me button, and integrations with Klaviyo and REVIEWS.io. The home page claims use by more than 14,000 Shopify brands, a vendor figure.
Terms list Feedback Group, Inc., Clearer.io UK Ltd. and Reviews.io 2020 GmbH among the affiliates. A group parent means roadmap and support decisions may not sit with the app team.
What stands out
Shopify search app
21-day trial
AppHub owned
Where it costs you
Plan amounts were not shown on the pages read, so cost is unknown before signing up.
Ownership by AppHub, LLC puts it in a group, which counts against it on independence.
Right for
Shopify stores wanting search and recommendations in one app with a long trial
Wrong for
Buyers wanting an independent single-product vendor or a non-Shopify stack
Not stated by the vendorPlan amounts not shown on pages read; 21-day trial
Spanish-language recommendations with retail media, CDP and loyalty modules
Ranked #10 of 14 in Best Product Recommendation Engines in 2026.
Pricing on requestElsewhere
The about page says the company was founded in 2013 and lists a Santiago, Chile contact address for Latin America. The site is in Spanish, with a catalogue that runs from retail media to loyalty to a customer data platform.
Controller details point to a group structure. The breadth suits a retailer buying many modules, but it makes the entry decision larger than for a single-purpose tool.
What stands out
Quoted proposal
Retail media
Live CDP
Where it costs you
The pricing page lists no amounts, so a budget needs a proposal first.
No country of establishment is stated, and the controller is a named group.
Right for
Retailers in Spanish-speaking markets wanting recommendations with retail media and loyalty
Wrong for
Buyers needing an English-language site and a self-service sign-up
Not stated by the vendorNot published; the pricing page asks for a tailored proposal
Developer-built recommendation models metered by request, with a free start
Ranked #11 of 14 in Best Product Recommendation Engines in 2026.
Free tierPublished pricingNorth America
The AI recommendations page describes collaborative filtering over user-item interactions, with Recommend Analytics tracking clicks, conversions and revenue.
The pricing page lists Free, Grow, Grow Plus and Elevate, with an extra thousand search requests costing less on Grow than on Grow Plus. The terms name Algolia, Inc. in Palo Alto for subscribers outside Europe. Plan the event tracking first, because the models retrain only from data you send.
What stands out
Recommend models
Metered usage
Developer APIs
Where it costs you
Three meters (Recommend requests, search requests, records) move the bill, so forecasting needs all three.
Elevate is quoted by volume, so the largest sites cannot see a price in advance.
Right for
Engineering teams adding recommendation carousels to a search platform already in place
Wrong for
Marketers wanting a no-code merchandising console and one fixed price
United StatesFree plan, then metered Grow plans by requests and records; custom Elevate
Commerce suite quoted on volume, with a proof of concept
Ranked #12 of 14 in Best Product Recommendation Engines in 2026.
Pricing on requestElsewhere
The pricing page says there are no rigid tiers and describes a Product Scalability package with a 99.99 percent uptime SLA, dedicated infrastructure and priority processing.
The company page lists offices in London, Paris, New York, Stockholm, Salzburg, Sydney, Helsinki and Berlin and over 140 team members, but names no headquarters country. A suite approach means buying search, bundles and email together if you want one vendor.
What stands out
GMV-based quote
Proof of concept
Eight offices
Where it costs you
No standard self-service trial exists, and the proof of concept is limited to qualified merchants.
Price follows turnover and traffic, so cost rises as the shop grows.
Right for
Mid-to-large retailers wanting a broad suite and a vendor-run pilot
Wrong for
Small shops needing self-service sign-up and a published price
Not stated by the vendorQuoted from traffic, turnover and modules; proof of concept instead of a trial
Retail-focused platform with recommendations and agents, no public price page
Ranked #13 of 14 in Best Product Recommendation Engines in 2026.
Pricing on requestNorth America
The about page says Eli Finkelshteyn and Dan McCormick founded the company in 2015, and the terms name Constructor.io, Inc. with a San Francisco copyright agent.
The home page frames products around an AI Shopping Agent and a Product Insights Agent rather than a plain recommendation widget. The audit has no strings attached, but it is an entry to a sales process, so keep the pilot scope narrow before expanding.
What stands out
Retail platform
Sales-led
AI agents
Where it costs you
The pricing address returns Page Not Found, so no cost can be estimated without sales.
The platform is framed around search experience, so a recommendations-only buyer may pay for more than needed.
Right for
Large retailers wanting a vendor-led programme with an audit as the first step
Wrong for
Small stores needing posted prices and a self-service start
United StatesNo public pricing page found; sales-led with a complimentary audit
Annual-scale commerce suite with product recommendations sold alongside search and marketing
Ranked #14 of 14 in Best Product Recommendation Engines in 2026.
Pricing on requestElsewhere
The pricing page says you pay only for the products you switch on and the volume you run. The privacy notice names Bloomreach Inc. as a controller outside the European Union, with Bloomreach B.V.
as its EU representative, and the copyright line names no parent. The home page says its Loomi layer connects to more than 175 tools. Expect a long evaluation before any recommendation widget goes live.
What stands out
Annual-scale
Modular suite
Demo only
Where it costs you
Only a demo request is offered, so a hands-on test needs a sales contact.
Quotes begin at an annual figure aimed at smaller brands, which rules out small shops.
Right for
Large retailers wanting recommendations bundled with search and marketing automation
Wrong for
Small stores that need monthly billing or a trial first
Not stated by the vendorRequest pricing; custom quote scaled by audience, products and volume
How to choose product recommendation engines software
A product recommendation engine turns shopper clicks, purchases and catalogue data into suggestions such as similar items, bundles and trending products, shown on a shop's pages, carts and emails. The differences that matter are rarely in the feature list, so this is the order we would work through them.
01
Start with pricing transparency
8 vendors on this page sell by quote, while 6 publish a pricing model you can read today: Recombee, Tweakwise, Barilliance, LimeSpot, Rebuy, Algolia.
02
Look at where the vendor sits
Check the vendor's home market (Elsewhere 8, Europe 3, North America 2, Middle East 1), because that is where subpoenas and data requests arrive first, whatever region holds your files.
Product recommendation engines: which unit sits on the invoice
Recommendation vendors disagree on what they count, and the unit decides how your bill behaves in a busy December. Recombee counts events, requests, users and catalogue size. Rebuy bands its price by average monthly Shopify orders. Algolia bills search requests, Recommend requests and records. Clerk.io prices each product by its own usage, such as recommendations or opened emails.
Nosto bases its quote on turnover and traffic, and LimeSpot's listed plan stops at a revenue ceiling. Barilliance sizes plans on traffic or annual sales. Write down your monthly orders, page views and catalogue size, then ask each vendor which of those moves the price. A seasonal retailer should also ask what happens in the peak month.
List monthly orders, sessions, SKUs and peak-month traffic first
Ask which meter triggers a tier change
Check whether a revenue or order cap applies
Testing a recommendation engine before you sign
Trial terms differ more than the feature lists do. Recombee gives an unlimited 30-day trial and then lets you stay on its free plan. Boost Commerce offers 21 days, LimeSpot 15 days and Segmentify 14 days, which it says is running within five. Nosto has no standard self-service trial and limits its pilot to qualified merchants, Constructor offers a complimentary audit, and Bloomreach shows a demo.
Luigi's Box and Clerk.io have trial buttons with no stated length. Barilliance names none. Since recommendation models learn from your own traffic, a two-week window may be too short to show uplift, so agree what you will measure and ask whether the models need a warm-up period.
Ask for the trial length in writing
Agree a measurable test before starting
Check what happens to your data when the trial ends
Shopify app, developer service or retail suite
The vendors in this market fall into three groups, and picking the wrong group costs months. LimeSpot, Rebuy and Boost Commerce describe Shopify stores, so they suit a merchant who wants installation without a developer. Recombee and Algolia publish API references and expect engineers to send events and render results.
Luigi's Box names seven platforms on its home page, which suits a group running different shop systems. Nosto, Bloomreach, Constructor and Retail Rocket sell wider suites with search, email, retail media or a customer data platform, so recommendations are one module in a bigger purchase. Decide first whether you have developer hours, then match the group, and only then compare prices inside it.
Confirm your shop platform appears on the vendor list
Count the developer hours you can spare
Decide whether you want a suite or a single module
Leaving a recommendation engine: terms, notice and your data
Exit is rarely on the pricing page, so ask. Barilliance charges monthly, binds you to nothing long and allows plan changes whenever you like. Recombee plans are monthly with a free plan to fall back on. Bloomreach frames its quotes as annual amounts, and Nosto sells scalability packages with an uptime SLA, both of which suggest committed terms to confirm.
None of the pages read for Tweakwise, Segmentify or Retail Rocket describe how interaction history is exported. Ask for the export format in writing, whether it includes the behavioural events or only the catalogue, and who owns the trained models. Test the answer before signing, because rebuilding event history elsewhere takes longer than moving the widgets.
Get notice period and renewal terms from the contract
Ask whether behavioural events can be exported
Confirm who owns the trained models after cancellation
What goes wrong most often when buying product recommendation engines software
Comparing starting figures without checking which of the modules (search, email, bundles) the normal shop needs to add.
Judging a recommendation widget after a short trial, before the model has learned from enough of your own traffic.
Choosing a Shopify-only app, then replatforming and finding no equivalent integration waiting for the new store.
Signing an annual suite for recommendations alone, then paying for search and marketing modules that stay switched off.
07
Frequently asked questions
6 answers
What is the best product recommendation engines in 2026?
Recombee leads our ranking of 14. Recombee, s.r.o. sells recommendations by the month and counts four things: interactions, recommendation requests, active users and catalogue items.
Paid tiers are published, and after the 30-day unlimited trial you can fall back to a free plan instead of paying. The pages read show an admin UI, an API reference and integration plugins, so expect engineering time for set-up rather than a one-click app.
Which product recommendation engines tools publish their pricing?
6 of the 14, with the pricing model each one publishes:
Recombee: Published monthly tiers by usage; free plan after a 30-day trial; top tier quoted.
Tweakwise: Free tier called Basic, then published from-prices for Advanced and Expert tiers.
Barilliance: Published entry price for personalisation, other plans by traffic or sales; monthly, no long contracts.
LimeSpot: Published flat monthly plan with all features; 15-day trial.
Algolia: Free plan, then metered Grow plans by requests and records; custom Elevate.
8 vendors do not print a price.
Is there a free product recommendation engines tool?
Recombee, Tweakwise, Algolia can be used for free, either as a free plan or a self-hosted edition.
Where are these product recommendation engines vendors established?
The headquarters are spread over 7 countries and 4 regions, Elsewhere 8, Europe 3, North America 2, Middle East 1:
Recombee: Czech Republic.
Tweakwise: Netherlands.
Luigi's Box: Slovakia.
Clerk.io: Denmark.
Barilliance: Israel.
LimeSpot: Not stated by the vendor.
Rebuy: Not stated by the vendor.
Segmentify: Not stated by the vendor.
Boost Commerce: Not stated by the vendor.
Retail Rocket: Not stated by the vendor.
Algolia: United States.
Nosto: Not stated by the vendor.
Constructor: United States.
Bloomreach: Not stated by the vendor.
What should you use instead of Recombee?
Tweakwise and Luigi's Box are the next two on this page. Tweakwise is for Merchandisers who want a tiered price ladder and a no-charge starting point; Luigi's Box is for Shops on one of seven named platforms wanting search and recommendations from one vendor.
Who should not buy Recombee?
Shop owners wanting a plugin installed without any engineering work. The pages read describe an API reference and a query language, so a team without a developer will struggle to use it..
Sell product recommendation engines and missing from the ranking? Send us a line on what the product does and who buys it. Whether it is added is our decision, though commercial arrangements can also affect it, and the terms of a listing sit on the software advice page.
