Video for startups, without the agency.

The problem with both options

When a startup needs video, there are usually two doors. Neither fits well.

A production agency quotes in weeks and thousands, because it has to — a shoot means a crew, a location, a cast and a schedule that all have to line up. The work is often excellent. It is also slow, and it prices most early-stage companies out of making video regularly rather than once.

A generation tool is instant and nearly free, and produces exactly what you would expect from something with no script and no director. It solves the wrong half of the problem. The expensive part of video was never the camera.

What sits between them

One producer, using AI as the camera rather than as the process.

Every project starts with a script and a storyboard, agreed before anything is generated, because a weak script produces a weak film no matter which model renders it. After that: generation, voiceover, edit, colour grade, delivery. Same craft as an agency, minus the twenty people and the truck.

You are talking to the person doing the work, not an account manager relaying notes to a team you never meet. Feedback goes straight into the edit.

What startups actually ask for

• Launch film. The one video that explains what you built and why it matters. Usually the hardest to write and the most valuable to get right.
• Product explainer. How the thing works, shown rather than listed.
• Feature announcements. Short pieces that ship alongside a release, in the same visual world as everything before them.
• Social cuts. Portrait versions framed for portrait, not cropped down from a landscape master.
• Fundraising and pitch material. The story, tightened, for a room that has seen forty decks that week.

Most of these are the same world, which is the point — the cast, the palette and the tone hold across all of them, so a year of releases reads as one company rather than a folder of unrelated clips.

Proof, not claims

MyThorneAI is an AI workforce-readiness platform with four distinct features. The obvious video walks through each one. Instead the brand film follows a single student, and each feature appears at the moment she needs it — so the viewer understands the product by watching it solve something.

Ten-plus videos came out of that project, landscape and portrait, with the same cast throughout. Read the full case study, or browse everything else.

What it costs to think about

There is no rate card here, because the same runtime can be a simple job or a hard one. What actually moves a quote:

• Runtime — more minutes means more shots, more generation, more edit.
• Scenes and distinct shots — a single continuous scene is cheaper than twelve locations.
• Complexity — character consistency across a series, lip-sync, localisation into another language, revision rounds.
• What gets delivered — a finished cut ready to publish, or raw generated clips for your own editor to assemble.

Say which of those apply up front and the quote reflects it. The full reasoning is in what AI video production costs.

Fair warnings

This is one person. That is the advantage and the limit. It means no handoffs, no account layer and direct answers — and it means there is a ceiling on how many projects run at once. Long-lead planning helps.

Cheap is not the point. If a quote is cheap because nobody is writing a script, that is what you will get. The saving should come from removed logistics, not removed craft.

How to start

Send what the product does, who needs to understand it, and roughly how long the video should be. If there are brand guidelines, send those too. You get a script and a storyboard back before anything is generated — so the first thing you approve is the idea, not a rendered video you have to react to.