Every second, read.
What’s seen, what’s said and what’s on screen in every video and image you run, each at the second it happens.
What’s seen, what’s said and what’s on screen in every video and image you run, each at the second it happens.
Published from Pluto, made in Ads Manager and kept in sync, or imported. Each file is read once it runs.
↳ See how it readsThe hook, the demonstration, the claim and what’s said, each a segment with its start and end.
# over MCP, illustrativesearch_creative_memory({ q: “the parka zipped to the chin”, limit: 20})# vire-down-parka-17s-16x9.mp4 at 0:05
Search the library by what an ad shows or says, ask the Agent, or call it from your own tools over MCP.
↳ Read the MCP docs
Built for the long cold.| 0:08 Seen | Warming his hands, in profile |
| 0:08 Moment | A claim, until 0:13 |
| 0:08 On screen | Built for the long cold. |
Each match says where it was found, on screen, said or seen, and when, so it can be checked against the ad.
↳ Read the Creative Memory guide# a claim, excerpt{ “value”: “Built for the long cold”, “evidence”: [{ “source”: “ocr”, “ref”: “ocr:1”, “startMs”: 8000, “endMs”: 13000, “quote”: “Built for the long cold.” }]}
Each segment keeps its start, its end, its words and where they came from. That’s what the Agent, Loops and your tools read.
Videos and images from all your ad accounts and platforms, understood the same way.
Your people, the Agent and Loops all draw on the same knowledge of what you’ve made.
Each ad you launch adds to it, so it knows a little more about your brand over time.
Creative Memory comes with Enterprise.Owners and admins choose what it reads.
When an ad goes live, every upload, or only when you ask.