Case study · Framewise
Side projectProject period2023
AI-assisted video analysis
Built an experience for uploading video and exploring insights from its content.
01
Why
A video holds much more than you can search for by watching it.
02
Decision
You upload a video and Azure Video Indexer extracts the insights. FrameWise shows them in a dashboard and on a page for each video, with a timeline: brands, emotions, faces, keywords, labels, mentions, places, sentiments and topics. The app refreshes its access token before each call. Quality comes from Jest for logic, Storybook for components and Playwright for end-to-end tests.
03
Trade-offs
It depends on Azure Video Indexer, which has its own limits and a fixed set of capabilities, so we can't get more out of it. Other options would have cost more, so we stayed with it.
04
Outcome
A live product with a dashboard of insights for each video.