How are numerical features supported in Superlinked? (text + numeric search)

[vemahendran]

I have been going through the Superlinked docs and examples to understand how numerical data is handled alongside text.

I can see that numerical features are supported in the schema (for example via NumberSpace), and a few examples combine text embeddings with structured fields like price, rating, etc. What I’m missing is a clearer picture of how this is used in practice beyond the basic demos.

A few questions for those who’ve tried this out:

• Have you used Superlinked with numerical data in a real use case or PoC?
• How do you typically handle scaling or normalization for numeric fields before putting them into a NumberSpace?
• When combining text and numeric data, how do you decide what should be part of vector similarity vs what should remain as filters?
• Any tuning tips, gotchas, or lessons learned when multiple spaces (text + numeric) are involved in ranking?

We’re currently evaluating Superlinked for a use case that needs joint search over text and numerical attributes, so real-world experiences or guidance would be super helpful.

Thanks in advance!