Hush raises $12M to scale privacy-preserving speech AI for UX research

AI · 4 min read

Hush's platform applies on-device anonymization, speaker disentanglement, and semantic tagging to user interviews and usability sessions so product teams can analyze patterns without storing raw audio. The company released a new SDK that can run locally on laptops and mobile devices to produce redact-ready transcripts for downstream analysis.

The $12 million Series A was led by Harbor Lane, with participation from design-focused angels and research labs. Hush plans to broaden language coverage, add sentiment and behavioral tagging models, and integrate with popular research repositories and analytics platforms.

Privacy advocates welcomed Hush's approach as a pragmatic middle ground between useful UX insights and regulatory demands, but warned teams must still design consent flows carefully. Hush emphasized built-in consent templates and audit logs as part of its enterprise offering.