❓ Problem

There are roughly 207 million content creators worldwide, around 50 million of them professional or semi-professional, inside a creator economy worth north of $250bn in 2026. All of them shoot constantly. Almost none of them can find anything.

Here is the bit nobody talks about. The bottleneck in content production stopped being the shooting years ago. Phones are cinema cameras, CapCut is free, and editing skills are a YouTube tutorial away. The bottleneck is retrieval. You know you filmed a decent sunrise shot at some point. You know there is a clip of your hands on a keyboard somewhere. You just have no idea which of 11,000 files it is, or which drive it lives on.

Past roughly 50 hours of footage, or 500-odd files, the coping mechanisms collapse. Folder structures rot the moment you shoot on a second device. Naming conventions survive exactly as long as your patience does, which is about a fortnight. Everything reverts to IMG_4523, and "that drone shot from the camping trip" becomes functionally lost. The footage still exists, it is just economically dead: you will never find it, so you go and shoot it again, or you buy stock, or you cut the shot entirely and the video is slightly worse.

Professional post houses solved this with media asset management systems, loggers and assistant editors. Adobe paid $1.275bn for Frame.io to own a slice of that workflow for teams. The solo creator, who has the same problem at a smaller scale and none of the budget, got nothing. They got a camera roll.

✅ Solution

An AI librarian that watches every clip you own and turns your camera roll into a searchable b-roll library.

• Connect the source, not the folder. Point it at Google Photos, iCloud, a hard drive or a Dropbox. It ingests everything, including the ten years of footage you had written off.
• It watches, then tags. A video understanding model goes through each clip and returns structured metadata: shot type (wide, close, drone, handheld), subject, setting, motion, time of day, lighting, mood. Ten years of noise becomes browsable folders like "golden hour walking", "hands typing" and "coffee pour".
• Search in plain English, or in script. Type what you want and get the clip. Better still, paste your whole script and it assembles a shortlist per line, then exports straight into CapCut, Premiere or Final Cut as a bin or a rough assembly.
• It learns your style and closes the gaps. After a few edits it knows what you actually cut with, and starts flagging what you are missing: "you have zero transition shots and nothing at night, here is a 10-minute shot list for your next walk."

The wedge is deliberately narrow. Not a media asset management platform. Not an editor. One job: make the footage you already own findable in under five seconds. Everything else, the shot lists, the auto-assemblies, the stock marketplace, comes after you have earned the library.

📊 Key Numbers

Market size

• 207 million content creators worldwide sit inside a creator economy valued at $250bn to $323bn in 2026, growing at 22-26% a year.
• The realistic top of funnel is the roughly 50 million professional and semi-professional creators who shoot regularly enough to have a library problem. At $15/mo that is a theoretical $9bn a year, which is the TAM you quote to a VC and nobody believes.
• The honest SAM is tighter: creators with a genuine archive, an editing workflow and revenue to justify a tool. Call it 3-5 million people globally, or a $540m-$900m annual pool at $15/mo.
• Worth noting only about 4% of creators earn over $100k a year. Price for the many, not the few. $15/mo is the right number precisely because most of this market is not rich.