Information overflow no more

We know how it is. The ever-growing pile of papers next to you on your desk is always tempting to ignore for just a while longer, as is the constant influx of new and ever-changing data in your company. For you to read through every line of every text in your project demands more hours of reading and analyzing than you necessarily have. To help you with the processing, Gavagai is at your service.

Multi-text summarization and topic modeling

The /stories and /topics APIs have been deprecated and are replaced by the Gavagai Explorer which also has an API (look towards the bottom of the page for the API documentation).

Multi-text summarization allows you to get the gist of your text collection without having to read through every single sentence. This way you can focus your analysis on the topics that matter. We expose two API calls for multi-text summarization – /stories and /topics.

/stories enables you to get a quick and informative overview of your provided document collection, where documents have been placed into different clusters – stories – along with similar documents. A story represents a summary of the document members based on their content and presents a short description of their content.

/topics significantly enhances the functionality of /stories and allows you to not only get an initial overview of their content, but also further explore different aspects or layers of the document collection.

Multi-dimensional tonality classification

Using Gavagai’s technology, we have developed an application for filtering the meaning of text in more dimensions than the mere positive and negative, which means that you can find out what the neutral words really say. Think of it like reading with a higher degree of resolution. As an example we can detect nuances of positivity like love or passion, or add sentiments like fear, hope or sadness. Or add other target categories that are important for your brand.

/tonality helps you extract these aspects of your documents. And since our system adapts to new language use constantly, you’re ensured that the data you receive always is up to date with new vocabulary and language structures.

Living Lexicon

Get to know the real side of your language. Discover the semantic context and explore unexpected uses for old and new words. This is the API backing Gavagai Lexicon. See for yourself at

The foundation of Gavagai is our semantic memories, that are constantly evolving and learning new words as they pop up in our collective minds. If you need to enhance your product with additional semantic information, /lexicon is the best tool for the job. Since we are constantly reading and analyzing everything we can get our hands on, from tweets to news articles, you don't need to worry about neologisms, informality, or wierd spellings.

Keyword extraction

Interested in knowing what entities are mentioned in your text collection? Perhaps you are looking into enhancing your search engine, or automatically tagging your blog posts with the most important topics mentioned. Or maybe you would rather like to create word clouds from your text collection that work better than the classic approach of simply ranking words by how often they appear.

/keywords extracts the most significant concepts found in your text collection. The algorithm has a direct connection to our semantic memories and uses the power of our living lexicon, Gavagai Living Lexicon, for constantly updating itself on new topics to include as keywords. We humans don’t need manually annotated data and retraining of algorithms to know who or what Charlie Hebdo is and why the words belong together, and neither should a machine.