Case Study: Seenit achieves pinpoint search of thousands of hours of user-generated video and cuts feature delivery from 12 weeks to 1 week with Couchbase

Seenit Gives customers powerful full-text search in the cloud

Seenit is a London-based startup that helps enterprises crowdsource authentic video from customers, fans, and employees—working with brands like Rolls-Royce, Red Bull F1, BT Sport, BBC, and Unilever. As user-generated libraries balloon to thousands of hours, Seenit faced the challenge of evaluating, tagging, and searching non-obvious video properties (visual objects, audio content, and sentiment) so producers could reliably find tiny, specific clips without manually watching everything.

Seenit built an automated pipeline on Google Cloud/TensorFlow to analyze videos frame-by-frame, convert results to JSON, and store them in Couchbase. Couchbase’s native JSON storage and Full Text Search (including fuzzy, wildcard, boosting, and faceting) made complex queries—like “people smiling in front of a pink sunset talking about family”—fast and precise, while its elastic scaling supported rapid growth. The platform cut feature delivery from 12 weeks to one, accelerated production workflows, and enabled brands to produce targeted, emotionally resonant stories from vast crowdsourced libraries.

Seenit

Dave Starling

Chief Technology Officer