An engineer's answer to the privacy wall.
Omna Labs is a bootstrapped, solo-founded company building the universal semantic layer between regulated enterprise data and AI. Started in New York. Built in Rust. Local-first by principle, not by feature.
Gaurav Jindal
Computer science by training. Years inside the financial sector, mainly on the data side — the part where every new column ships with a compliance review and every model launch hits a privacy wall.
That experience is the reason Omna exists. The pattern is the same in every regulated industry: the smartest models on earth, blocked from the most valuable data on earth, because nobody wants to be the engineer whose pipeline leaked a customer record.
I ran out of tokens. Everyone is running out of tokens.
Every company on earth is trying to push the same documents into the same AI models — and every one of them is hitting the same two walls. The privacy wall: legal won't let raw records leave the building. The token wall: even when legal gives the green light, the bill is enormous and the context window fills up before the model has seen anything that matters.
AI is smart but blind. Your data is rich but invisible to AI. Omna is the bridge.
I lived this pain inside banks. Four years writing brittle regex pipelines to scrub sensitive columns before any of it could touch a model. Every new schema meant a new edge case, a new on-call page, a new audit finding. The tooling was supposed to exist. It didn't.
Omna is what should have existed all along — a .search() and .mask_pii() that are first-class operators on a Polars DataFrame, with a Rust kernel under the hood and zero network calls by design. Started as an open-source Python library so any engineer could install it and feel the difference in five minutes. Grew into a Mac app and a browser extension so anyone — not just engineers — could push files into Claude, ChatGPT, Gemini, Perplexity, or DeepSeek without uploading their privacy along with them.
We're not trying to replace AI. We're trying to make AI usable on the 90% of data that AI currently can't see. That's the wedge. From Polars today, to every format and every model by 2027.
Four principles. Everything we ship is downstream of these.
"They cannot see your data until you show it to them in exactly the right way."
"Omna is to AI models what a good search engine is to websites."
"AI is smart but blind. Your data is rich but invisible to AI. Omna is the bridge."
"Google didn't kill websites — it made them more findable, and therefore more valuable. Omna doesn't replace AI. It makes AI usable on data that was previously unreachable."
Build the universal semantic layer with us.
Solo founder, no investors, no board — just engineers who care about regulated data and want to ship the tools that should have existed all along. If you've felt this pain too, we'd like to hear from you.
Backing the bridge between enterprise data and AI?
We're bootstrapped today — and intentionally building product before raising. If you invest in infrastructure for regulated industries, we'd love to keep you in the loop.