A search experience that actually converts

9% increase in Add to Cart rate
12% uplift in Click Through rate
80% decline in "No Results" pages
See it in action

Twiggle applies natural language understanding to e-commerce search

Our natural language processing technology coupled with our proprietary ontology understands and structures both product listings and shopper queries.

Leveraging this intelligence, we’ve developed an API that helps retailers boost their search relevance and connect shoppers to the products they’re looking to buy.

Connect shoppers to what they’re looking to buy

Our solution leverages advanced NLP technology coupled with a proprietary rich ontology to enhance relevance and get shoppers from search to cart, faster.

Twiggle understands shopper queries

Twiggle understands product listings

Twiggle complements your existing search engine

We integrate seamlessly with any search engine to produce best-in-class search results while preserving existing business rules and search logic.

Natural language search

The backbone of delivering a good search experience lies in the search engine’s ability to understand the way shoppers express themselves in natural language.

To understand shoppers’ intent, search engines have to recognize & support these ways of expression, which we call linguistic and inference phenomena.

Category Search

Shoppers can enter a general search term and retrieve lower level product results for all items in the given category. For example, a search for furniture will display results that include chairs.

Product Type Inference

Twiggle understands the product even if the shopper doesn’t explicitly specify the product name, i.e. black Vans.


Same same, different name. Shoppers can mean the same thing but refer to it in different ways.


Customers can describe the product based on what they don’t want, i.e. sleeveless dress. Your search engine can now understand “not, without, -less,”.

Numeric Values & Ranges

Shoppers can search by size, length, dimension and height. Twiggle can also grasp value ranges and return relevant result such as “baby clothes, 1-6 months”.


Shoppers can use qualifiers to specify various conditions that the products should meet for them to purchase, i.e. expensive shirt or cheap laptop.


Twiggle can correctly differentiate the main product type from its components i.e. dress (main product) with sleeves (components) and can link the correct attributes to right tokens in the query.

Attribute Conjunction/Disjunction

Some words stick together- like “black and blue” or “faux leather”. Shoppers can describe the product they’re looking for in detail and we’ll make sense of it for your search engine.


Shoppers often are interested in buying accessories compatible for products they already own, like a charger for iPhone X. We can identify the product the shopper is actually looking to buy.


Some products go hand in hand. Twiggle identifies products that belong together to enable customers purchase a set of products- better for business and easier for customers.

Natural language utilizations

Shoppers can shop for products like they would in-store, with the nuances and specifications of natural of language, and receive relevant product results.


Adept at auto correct. Our spell checker covers the world of consumer products and is built from real user queries. Our dictionary is always up to date with new trends and brands.

Let’s shop



Optimized search queries


Indexed, structured product listings


Processed queries per second

Twiggle’s Secret Sauce

Our ontology digitizes the world of retail.

Our teams have built the first knowledge model geared exclusively for the world of consumer products.

The ontology holds over 10,000 product types, 800,000 attributes and millions of attribute values.

With this knowledge, we can assign meaning to each word in the query string and capture shopper intent.

It also allows us to structure and normalize product listings so we can match the products in your catalog to the query.

Learn more