AI video briefs that start where the last project ended

Think about the first hour of a new AI video project. Most of it is not making anything. It is writing the brief again: the product rules, the client’s colours, the lighting the team likes, the format the client always asks for, the things that went wrong last time.

None of that is new information. Your team already knows it. The tool just does not.

Creative memory is how Blooper carries that knowledge from one project to the next, so a brief can start where the last project ended.

A one-line brief that already knows things

On a project with memory, a brief can be short. “Same world, ten seconds, horizontal” is enough to start, because the rest is already on your team’s memory page:

• your own habits, such as approving stills before anything moves,
• the team’s habits, such as “the product is never cropped”,
• what the client has settled, such as their house colour,
• and the rules: the aspect ratio, model or resolution your team or client has decided.

Before the agent plans a shot, it reads the notes that apply to this project, this client and you. Rules are checked on every generation the agent makes, before it is priced. Under your message, a small line shows how many team memories were in play, with a picture of each, so you can see what the first frames were built on.

You still direct the work. If today’s brief asks for something different, today’s brief wins. Memory saves you from repeating yourself. It does not stop you changing your mind.

Corrections you only make once

Most of the time a team spends on AI output goes into corrections. “Keep the label to camera.” “Less saturated, deeper shadows.” “Not on the lips.” Without memory, you make the same corrections on every project, because every chat starts blank.

Blooper watches for corrections: a prompt you edit before approving, a message you type while a run is going, or a change you ask for right after a shot comes back. When a correction looks like something that should hold next time, Blooper writes it as a short note and puts it in your Inbox. Accept it and it applies from then on. Dismiss it and it is gone.

Repeated settings are caught too. Ask for 9:16 for the same client in three messages across two chats within two weeks, and Blooper suggests “Usually 9:16 for this client”. Accept it, and if it should never vary, a team admin can make 9:16 a rule for that client, so nobody types it again.

Team rules anyone can read

A lot of a team’s know-how lives in one or two people’s heads. The senior designer knows the client hates fill light. The account lead knows the logo never sits on the left. When they are busy or away, the work slows down or goes wrong.

On the Team memory page, that know-how is written down where everyone on the team can read it, as tiles with pictures:

• Notes show what the agent takes into account, with the image or clip each one is about.
• Rules show what is enforced, for the whole team or for one client.
• References hold the images your team keeps on purpose.
• Search and an “applies to” filter narrow it to one client or one project.

Each memory says whose it is, too: Personal (only you), Team (shared), or Blooper (general notes about the tools, which you can mute in a project).

Because notes are short and in plain words, the page doubles as a style guide your team did not have to sit down and write.

Onboarding a new teammate

When someone joins your team, the team’s shared memory reaches their chats from their first project. Shared notes go to whoever is on the team right now, so there is nothing to copy or send. They do not need the senior designer to walk them through the client’s preferences; the agent already works with them, and the memory page shows them why.

Guests are different. A guest invited to comment on a project does not get your team’s memory in their chats.

And when someone leaves, what they taught the team stays. A note a member wrote for the team outlives their account. The knowledge does not walk out of the door with the person.

Onboarding a new client

For a new client, start with what they give you. A team admin can import the client’s brand book, as a PDF with selectable text, a Word file, or a plain text or Markdown file, up to 100,000 characters. Blooper reads it and proposes notes and rules. Nothing is ticked in advance; you confirm each one, and choose whether a rule applies to the whole team or to that client.

From then on, every project for that client starts with those rules and notes. What you settle on the first project, such as the house colour the client asked for in their notes, can be kept as a note for that client, and it carries into the second. A client’s memory stays with that client and does not reach your other clients’ work.

The time it gives back

We have not measured how many hours memory saves a team, and we will not invent a number. What we can say is where the time goes today, and which parts memory removes:

• the brief you rewrite at the start of every project,
• the corrections you make again because the last chat forgot them,
• the walkthrough a new teammate needs before their work matches the team’s,
• the rounds of review that catch a mistake someone already caught last month.

If your team spends its first hour on every project re-explaining the brand, that is the hour memory is built to give back.

Where it is today

Creative memory is in early access. It runs on Blooper’s own team first, and a team admin switches it on for their team when it reaches them. If you want it on yours, tell us.

How memory works is in the first post of this series, and what it saves in credits is in the second.

Start a project in Blooper, or see pricing.