Origin
The problem started with our own voice memos
Ruby spent years working directly with B2B founders on audience-building and content strategy. The pattern she kept seeing was the same: a founder with a sharp observation from a client call, a voice note sitting unread in their phone, a half-formed opinion that never made it to a post. The ideas were there. The time to turn them into something publishable was not.
She tried the obvious solutions alongside her clients. Generic AI tools gave them posts that sounded technically correct but nothing like the person who recorded the note. Hiring a ghostwriter produced clean copy that lacked the actual texture of how the founder thought. The personal voice, which is the reason a founder's post gets read over a brand post, was gone.
Daniel had spent years building NLP and voice-processing pipelines for marketing platforms. When Ruby brought him the problem in early 2025, the technical path was clear: the missing piece was a system that learned style, not just processed text. They started building Curioeus that same month. The question they kept coming back to was: what if the conversion did not just clean up the words, but actually learned the way the person thinks?
That question shapes everything we build. Voice first. Precision second. Speed third. In that order.
2025
Founded in Singapore
50
Early access cohort seats
<15m
Voice note to published post