API docs

FAQ

Cross-selling and upselling in e-commerce: frequently asked questions

A short guide to what crosello.io is, how cross-sell and upsell recommendations work, and how to connect a store.

crosello.io at a glance

What it is
A cross-sell and upsell product recommendation engine available through an API.
Who it is for
Marketing automation platform providers and their customers — online stores.
Input
The store’s product feed as a CSV file (required: product ID and name).
Output
Two lists for every product: cross-sell (accessories and consumables) and upsell (pricier substitutes and products used together).
Integration
HTTP/JSON API with an API key and no request limit.
Time to lists
Usually 10–40 minutes after the feed is uploaded; after that API responses are served without delay.
Price
From PLN 1 per month for 1,000 products; the final price depends on catalog size and contract length.
Cost vs. an AI agent
About 19× cheaper than matching the same 5,000 products with an AI agent (measured token cost).
Availability
Available through the API; plugins for WooCommerce, PrestaShop, Shopify and IdoSell are planned.

What is crosello.io?

crosello.io is a product recommendation engine available through an API. From a store’s product feed it builds two lists for every product: cross-sell (accessories and consumables that fit it) and upsell (pricier substitutes and products used together). Matches come from analyzing the relationships between products in the store’s catalog, not from hand-written rules.

What is the difference between cross-selling and upselling in e-commerce?

Cross-selling means suggesting products that complement what a customer is viewing or buying, such as toner for a printer, a water filter for a coffee machine or a case for a smartphone. Upselling means suggesting a pricier version of the same kind of product, or a product used together with it. Both raise order value: cross-selling adds something new to the basket, while upselling moves the customer up from what they were already considering.

In crosello.io they correspond to the cross_sell and upsell lists. The upsell list holds pricier substitutes (price ×1.1–1.8 of the product) and products used together.

How does the crosello.io recommendation engine work?

You upload a product feed (CSV). crosello.io analyzes the catalog: it builds a relationship matrix between categories, reads product cards (brand, series, model, system) and computes cross-sell and upsell lists for every product. Language models work once, while the feed is analyzed; the lists are computed once per run and served through the API without delay, so no model call is needed for each request.

What data does crosello.io need and how long does feed processing take?

A CSV file with the product catalog is enough. Product ID and name are required; categories, brand, description, specification, price and stock are recommended. Columns are detected automatically. Processing a feed usually takes 10 to 40 minutes, after which the lists are available through the API.

How do I get cross-sell and upsell recommendations through the API?

You send the ID of one product or several products (for example a customer’s recent purchases) and receive cross_sell and upsell lists. Every item carries a relation (for example accessory_for, consumable_for, substitute_of) and a proof of fit (evidence). Requests are authenticated with an API key; requests are not rate-limited.

POST /v1/feeds/{feed_id}/recommendations
{"product_ids": ["4411", "4418", "2203"]}

Which marketing automation platforms and stores does crosello.io work with?

crosello.io is built for marketing automation providers and their customers — online stores. Integration means calling the API (HTTP, JSON) with the IDs of products a customer viewed or bought; the platform puts the returned products into an email or newsletter. Plugins for WooCommerce, PrestaShop, Shopify and IdoSell are planned but not yet available.

How much does crosello.io cost?

Pricing depends on catalog size and contract length. From PLN 1 per month for every 1,000 products (typically PLN 1–2), with no limit on API requests. For a store with 5,000 products that is from PLN 5 a month.

How is crosello.io different from matching products with an AI agent?

An AI agent analyzing thousands of product descriptions consumes an enormous number of tokens. For a catalog of 5,000 products the measured token cost of an AI agent (Claude Opus 5.5 via the API, medium setting) was PLN 97, while crosello.io costs from PLN 5 a month — about 19× cheaper — with no limit on API requests.

The lists are computed once and served from ready data, so API responses do not depend on model run time.

Are the recommendations based on customer behavior?

A match is decided by the relationship between products — compatibility, function and need — not by what other customers bought. Purchase and browsing history is an additional signal: a request can include viewed categories (viewed_categories), which raise items from those categories (score ×1.5), and the IDs of purchased products (product_ids) are the source of the recommendations.

Which stores and industries is crosello.io suited for?

Catalogs in which products are linked by compatibility: accessories for devices, consumables and products used together — for example consumer electronics, home appliances and TV/audio, smartphone accessories, coffee equipment or printers with toner. The engine has no built-in knowledge of any industry; the store context is derived from its feed.

How do I get started and try crosello.io on my own catalog?

Use the “Add a feed and try it” button: you create an account in the panel, upload a CSV file with your catalog and see the cross-sell and upsell lists for your products. The API documentation describes keys, endpoints and sample requests. Questions about integration, pricing and investment cooperation can be sent through the contact form.

Didn’t find your answer?

Try crosello.io on your own catalog or write to us — we will answer questions about integration, pricing and cooperation.

Glossary: cross-selling, upselling and product recommendations

E-commerce and marketing automation terms as used in crosello.io.

Cross-selling

Also: cross-sell, cross sell

A sales technique of offering a customer an additional product or service that complements what they are viewing or buying — for example a case for a smartphone, a filter for a coffee machine or toner for a printer. In e-commerce, cross-sell recommendations appear on product pages, in the cart, in post-purchase emails and in newsletters. Wikipedia
Upselling

Also: up-selling, upsell, up-sell

A sales technique of offering a pricier or better-equipped version of the product the customer is considering. In crosello.io the upsell list holds pricier substitutes of the same kind (price ×1.1–1.8) and products used together. Wikipedia
Product recommendation engine

Also: recommender system, recommendation engine

Software that chooses which products to show a customer, based on data about products, customers or both. crosello.io relies on relationships between products (compatibility, function, need), not on the behavior of other customers. Wikipedia
Product feed

Also: product data feed, catalog file

A file listing a store’s products and their attributes: ID, name, categories, brand, description, specification, price, stock. crosello.io accepts a CSV feed and analyzes the catalog from it. Wikipedia
Marketing automation

Also: marketing automation platform

Automating marketing activity: sending emails, SMS and notifications triggered by customer behavior, such as a purchase, an abandoned cart or a viewed product. A marketing automation platform can fetch products to insert into a message from the crosello.io API. Wikipedia
Recommendation API

Also: product recommendation API

An interface (HTTP, JSON) through which an external system sends product IDs and receives lists of recommendations. In crosello.io: a GET request for one product or a POST request for a set of products, for example a customer’s recent purchases. Wikipedia
Product relations

Also: accessory_for, consumable_for, complementary_to, substitute_of

crosello.io distinguishes four relations: accessory_for (an accessory for a device), consumable_for (a consumable), complementary_to (products used together) and substitute_of (a substitute of the same kind). The first two form the cross-sell list, the last two the upsell list.
Proof of fit (evidence)

Also: evidence, match confidence

A field on every recommendation: confirmed (fit confirmed on product cards), probable, category (from the category, without a fit declaration on the card), similarity (description similarity) or weak.
Average order value (AOV)

Also: AOV

Revenue divided by the number of orders. Cross-selling and upselling are among the most common ways to raise it.
Large language model (LLM)

Also: language model, AI agent

An artificial intelligence model that understands and generates text, used among other things to analyze product descriptions. In crosello.io language models work once, while the feed is analyzed; API responses do not require a model call.