vidlink annotates the moments that matter inside a video.
like writing in the margins of a book — but the margins are the frame, and the notes are links.
paste a video. ai watches it and finds every product, person, place, reference. each one becomes a clickable link, timed to the exact moment it appears on screen.
the look is a field notebook drawn over the video. a hand-drawn outline traces the real object. a thin line points to its name. minimalist, scientific, playful.
not “interactive video.” not buttons or hotspots. just the moments that sparked something, marked and linked.
this started with a tweet in 2009. a video producer asking a simple question: how do you add links on videos? nobody had a good answer. same idea, four names, seventeen years.
"thinking abut embedded links on videos...any good idea or provider?"
first attempt. an image trigger called who shot that. it recognized what was in the frame.
"today I got my prototype. excited." — looking for beta testers. then life happened. the developer got a job. production took over.
apple buys shazam for $400m. same instinct: the device recognizes what you see.
"this is why I enjoy xray by prime." — the idea kept surfacing. annotate the frame. link the moment.
ai made it possible to build alone. one paste, one video, every link found. vidlink shipped.
the idea was always there. the ability to build it was not.
in 2014 the prototype worked — the image trigger recognized objects in the frame. but the developer moved on, and production work paid the bills. the idea waited.
then ai tools let one person build what used to take a team. a guy with zero experience in tech, building an app by asking. vidlink is what came out.
other tools make you place buttons on a video by hand. drag a hotspot, type a url, set a timecode. that works for ten links. it does not work for a thousand videos.
vidlink's ai watches the video and finds the links itself. it sees the sneakers, the camera, the location, the song. it knows what frame they appear in and where they point.
nobody else does this. the nearest competitors are small shops selling manual hotspot tools. none of them have ai that finds the links.
computer vision keeps getting better. the frame will recognize every object, every brand, every face — by itself. no manual work at all.
the dream from 2009 was always the same: you watch a video, you see something, you tap it, you go. that simple.
built by ignasi vargas. video producer, barcelona. been thinking about this since 2009.