We continue to improve TextKit adding state-of-the-art functionalities to it using in real-world use cases!

2Chat​

As the makers of TextKit, we at Nagence have launched a new product in the e-commerce industry called 2Chat, which focuses on automating Customer Support for Shopify stores. APIs like Emotion and Sentiment detectors came really in handy here to help us create a product that actually adds value to Shopify merchants.

What's new also is that 2Chat helped us find other pain points in the product we are building which ended up leading to 2 new launches. Read on!

New launches​

World cloud generator​

Which we are using to create reports of the most commonly mentioned topics that can help users understand in a single look what customers are saying among hundreds of conversations.

Link to the documentation

Sentence similarity detector​

Created specifically to help filter out chats that are referring to the same topic but in a different way. This endpoint also proved to be really useful to compare topic descriptions and find duplicates among them without having to recur to humans reading hundreds of chats manually.

Link to the documentation

We are pleased to announce our Zapier integration is now live in Zapier's marketplace. Go give it a try!

Together with our RapidAPI integration that has been live for a few weeks already, this integration is exciting news for our users wanting to use the power on Natural Language Processing in places where we don't have direct reach yet.

🎉One exciting announcement we wanted to make is that we added two new APIs to our list of Detectors 🎉.

Names API​

It takes your input text and looks for first and last names returning the ones that were a match.

The use-cases we had in mind when we developed this API come from our experience running automated marketing campaigns where we wanted to determine if the user part of an email address contained a name, nickname, or an unknown word.

For example, given the email address john.smith@example.com you can extract the user part of it which will return john.smith. With that, you can then search for common separators people use for their username, such as dots, hyphens, and underscores to split the text into single words that you can later pass to the Names Detector API.

Finally, the API will return John as the first name, Smith as the last name, and with that information, your automated email campaigns can greet the recipient with a "Hi, John!" instead of the random "Hi!" or "Hello there" that often indicate to the recipient they are being cold-emailed.

Link to the documentation

Gender API​

Similarly, this API takes as input a person's name and tries to determine, using a database of over 200k examples, if the name is indicative of a male or female gender.

As a potential use-case, you can utilize this API to dynamically change a text where the need of using pronouns is present or in the case of languages other than English, it can come in handy with nouns that are written differently according to the gender of the subject. For example, in Spanish, Bienvenido, José, and Bienvenida, Martha.

Link to the documentation