# The AI Assistant: One Sentence In, a Finished Video Out

> Ask for a video and the assistant writes the brief and the dialogue, then shows the price in credits before anything spends.

Canonical page: https://www.socialfaktory.com/blog/ai-assistant-for-social-video
Published: 2026-09-06
Language: en

Until now, making a video here meant opening the composer and filling it in yourself. There is nothing wrong with that. The composer is where every knob lives, and the knobs are the point. But most days you already know what you want, and you just want to say it.

So now you can. The AI assistant sits in a bubble on every page of your account, and it will take one sentence and hand back a finished brief, the words your presenter speaks, and the price in credits before a single one is spent.

## Where it lives

A draggable bubble on every page, and a full page of its own when you want more room. Drag it out of the way and it stays where you put it. It reads the page you are on, so you rarely have to explain yourself twice.

## Path one: say it, price it, generate it

You are on your dashboard. You type:

> Make a video for my new serum.

It checks your brand and your products first, so it does not ask you what the serum is. That is already in your account. A few seconds later it comes back with a card:

- **Video · 20s · 9:16**
- **Brief:** Close on the bottle, then the morning after
- **Says:** "Three drops at night. That is the whole routine."
- **12 credits**
- Generate video · Not now

That card is the submit button. Nothing has been spent yet. Tap **Generate video** and the render starts, and the assistant hands you a link to the page where you can watch it run. Tap **Not now** and the whole exchange has cost you a fraction of a credit.

Two things are worth noticing. First, it wrote two different things: the brief, which is what happens on screen, and the dialogue, which is what the presenter actually says. Those are separate fields, and they are the two hardest lines to write. Second, the number was on the card before you touched anything.

## Path two: watch it fill the composer

The composer has not gone anywhere, and it is still the better path when you want to steer. What changed is that you no longer have to start from an empty form.

Open the composer and press **Write it with the assistant**. The panel collapses to a small pill so it is not sitting on top of the form, and then the fields fill in front of you, typed one at a time rather than pasted in a lump. The step holding the field being written scrolls into view first, so you are always looking at the thing that is changing. The speech length indicator ticks along while the dialogue types, so you can watch the words fit the duration you picked.

When it finishes, it tells you what it set, what is still missing, and the one thing to do next, with the price. Something like: brief, dialogue and format are set, pick an actor, then hit Generate, 12 credits.

Then you press Generate. Not the assistant. You. In the composer, the composer's own button is always the submit, and it always will be.

## Why the price is on the card

Plenty of tools spend your money and tell you afterwards. We are not going to do that, so the rules here are boring on purpose.

Anything that generates shows its full price in credits on a card, before it runs, and it does not run until you tap the button. The conversation itself costs credits too, charged on what each answer actually used rather than a flat fee per message, which means asking a question is cheap and making a video is not, and the card says which one you are about to do. If a generation fails, the credits come back automatically.

## What it will not do

This is the short list, and it is deliberate.

It cannot publish. It cannot schedule. It cannot delete anything. It cannot look at another brand's work, and it cannot reach anything outside your account. Ask it to write you a shell script and it will decline in one sentence and point you back at your content.

That is not a gap waiting to be filled. An assistant that can do anything is an assistant you have to supervise, and the whole point of this one is that you do not. Finished videos wait in your library, and nothing goes live until you schedule the post yourself.

## One more thing: drop in a picture

If you have a reference, a product shot, a frame you liked from someone else's ad, drag it straight into the chat. The assistant reads it, and the image lands in your media library as an asset you can use anywhere else in the app. You never have to upload it twice.

## Try it on your next brief

The next time you catch yourself opening the composer and staring at an empty brief field, click the bubble instead and say the sentence out loud in text. Read the card. If the number and the words look right, tap the button. If they do not, tell it what to change and read the next card.

That is the whole product now: you describe, it proposes with a price, you decide.

[Get started](/register), or read [how SocialFaktory works](/blog/how-socialfaktory-works) first.

## Questions

### Does the assistant post to my channels?

No. It cannot publish or schedule anything. What it makes lands in your library, and nothing goes live until you schedule the post yourself.

### What does the assistant cost to use?

It runs on the same credits as everything else, charged on what each answer actually uses. Anything that makes a video shows its full price in credits on a card first, and nothing runs until you tap Generate.

### Can I give it a picture to work from?

Yes. Drag a PNG, JPEG or WebP into the chat. The assistant reads it, and the image lands in your media library as an asset you can reuse anywhere in the app.

## More detail

- All articles: https://www.socialfaktory.com/blog.md
- What SocialFaktory does: https://www.socialfaktory.com/features.md
- Pricing: https://www.socialfaktory.com/plans.md