An clever content indexing solution from Nuclia

Nuclia, a young Spanish software company based in Barcelona, Spain, has participated for the first time to the recent 47th edition of The IT Press Tour. And it was a great surprise as the solution really rocks. Let’s dig a bit.

First the company, Nuclia was founded in 2019, has around 25 employees, and received so far 5.5 million euros as a seed round from Crane Ventures Partners and Elaia. The founding team already collaborated together in past project like Iskra.cat, Intranetum.com or Onna.com.

The challenge they address is well known and huge so it invites several players to think about how to index the content of a large volume of unstructured data. As we all know indexing metadata is pretty easy, well mastered by plenty of solution, but the real grail is the content. The recent Coldago Map for Unstructured Data Management unveiled 20 players coming from the storage industry but only a very limited number of them know how to deal with content. One of these is Data Dynamics that understood that need very early and decided to acquire the Indian company Infintus Innovations Pvt. Ltd in 2019, 3 years ago.

The problem can be summarized in a few lines:

• The volume of data explodes especially the unstructured data type, it is true for the number of files, the nature of them and their size,
• The format of files themselves need the right level of interpretation. Again metadata, extended attributes of files can be extracted via classic file system level tools or API calls but the internal data is another story. And let’s consider the language dimension associated with a large variety of data sources, these create a certain complexity.
• And keywords have hit a wall, it is not enough to manage content with this approach.
• So clearly the market requires a new level of solutions that are able to discover all type of file content to enable a new level and more comprehensive knowledge to build a real analytics landscape you can navigate into.

So Nuclia has jumped into this mission with new ideas, talents and expertise. The key decision they made is to offer AI-based search as-a-Service with a very simple way to submit unstructured data to the index engine without any special code to generate. The architecture, shown below, is a multi step process controlled by the Nuclia Desktop, and SDK or a REST API. The other element central and fundamental in Nuclia’s approach is the design of a dedicated database, the famous open source vectorized NucliaDB, as they didn’t find anyone on the market suitable for their need.

The service exposes an easy 6 steps workflows with:

1. Selection of any data source,
2. with any language for any kind of data,
3. then an extraction from the source whatever the type text, audio or video,
4. establish insights which is more than a keyword being a trend or a meta topic if I can say that,
5. then creates vectors
6. to insert info as records in the NucliaDB.

And the beauty of this is the Google-like search user experience with just one field to access content.

So far the company has identified and chosen some use cases more vertical ones but not yet data management solutions like the ones mentioned in the last Coldago Map 2022 for Unstructured Data Management.

I invite you to try the solution, I tried it and it delivers very interesting results in a very simple easy experience, so just go to https://nuclia.com/sign-up/.