Nuclia makes the unsearchable searchable, raises $5.4M, and open sources NucliaDB
Almost a year ago, we were introduced to Eudald and Ramon, founders of what was then called Flaps (think ailerons), by our friends at Aldea. Morgane and I took the first call. They pitched an enterprise knowledge management product. For those who don’t know the space, it’s stupendously boring and not what you would think of as a growth market. On top of that, it’s hard to build anything that doesn’t just devolve into people pinging deep links to each other.
Then they did the demo.
Hold. The. Damn. Phone.
It was so much more than KMS. If you look closely, here are some things you’ll see:
- Searching turns up structured and unstructured data, like video
- In each file, you are taken to the block of text or time (?!) with the relevant content
- Relevance is not just keyword-based, but conceptual, aka semantic
- There’s a sidebar of related content
So, what they had built was search against unstructured data across media types that’s both keyword and semantic and does relations. 🤯
The next natural questions were:
- Is this for real?
- Are there tricks that make it look like something it’s?
- How does it work?
- What are the limitations?
- What’s the stack of technologies underneath?
- Can we strip the KMS UI off it and use it for other applications?
Which led to a discussion about why KMS, what Eudald and Ramon were really setting out to build and for who, what they really cared about, and what dent they wanted to make in the future.
Fast forward to Nuclia, which takes the core tech behind what they showed us in that demo and provides it as an API that can be plugged into any app or service to provide keyword and semantic search against unstructured, multi-media, and multi-lingual source material. It’s bloody magic. ✨
Nuclia makes the unsearchable searchable.
Yesterday, they released the API service in private beta. Sign up at nuclia.com to get access today!
They also open sourced NucliaDB, the first vector database designed for unstructured data. You can bring your own vectorization and normalization—while NucliaDB provides storage, indexing, and querying. You get the same vector, keyword, and relational search that powers Nuclia. Find it on Github at github.com/nuclia/nucliadb.
Last but not least, we’re very proud to lead their $5.4M Seed round alongside Elaia . 😃🥳🚀