That database sounds like scam created by junior programmers that invest the major part of their VC budget money into marketing instead of engineering. I don't believe that they are even fully aware of all functionality of ES that they are comparing to.
I recently did a quick research about fulltext engines and found it to be very frustrating.
The most promising with a comment /note:
ElasticSearch - good, but huge and strange licensing
Meilisearch - my favorite, nothing to complain so far
QuickWit - good but *nix only
TypeSense - good but *nix only
Manticore - untested but looks interestung
ZincSearch - seems abandoned, issues with ES API compatibility
Solr - interesting but a bit old fashioned
Sphinx - DB based, old fashioned
With DB I meant SQL / DBMS based. Most DBMS have support for fulltext, while mostly nowhere near the performance or featurette of specialized systems.
Elastic, solr and others are based on lucene, and while old fashioned is nothing Bad, it may lack features like extended monitoring, distributed services or optimizations for modern Hardware.
For Vespa there's a managed version hosted by the Vespa company in their cloud environment, and then the open source version is easily run locally or in any environment of your choosing. It takes some attention to detail, but it's quite flexible. I have a long running single node instance on an Intel NUC, but I've also run more complex cluster variations across different cloud environments.
+1 for replacing ES with ParadeDB. I did this for one of my projects recently and it's working just fine for my use case (FTS), with much less overhead. The DXP could use some improvement, but I think they'll get there. Good to see ParadeDB getting more traction.
There's a pre-existing aws-s3 extension allowing for s3 object connectivity. A duckDB extension is very different from this in scope. Also, s3 file access seems to be referred to as data lake access? Confused.