Who is this actually for?

Mostly direct-to-consumer brands selling a physical product online — skincare, supplements, apparel, home, pet, beverage, fitness. That’s what the pipeline is built around: we work from your reviews, your product photography, and your TikTok, Reels and paid results. We do take work outside it — there’s software work in the grid above — but if you’re B2B, an app, a service, or a local business, say so at intake and we’ll tell you honestly whether we’re the right shop before you pay anything.

Is this organic content or paid ads?

Both, and that’s the point of buying it in one place. The same angle gets cut for your feed and for your ad account, and what wins in one tells us what to make for the other. Most brands run these as two disconnected workstreams and pay twice for the privilege.

Who owns the content?

You do, outright, the day it lands. Full commercial rights on every platform, forever, with no expiry, no renewal fee, and nothing to re-license. Worth being precise about, because a creator deal is a licence, not a sale — those rights typically run out after 6 or 12 months, and extending them costs you again. Ours don’t run out, and nothing you have bought can be taken back.

Why so few videos? Everyone else offers more.

Because more stopped being the hard part. A hundred videos a month from a pipeline that isn’t reading your results is a hundred videos you have to sort through yourself. We’d rather send twelve we can defend and tell you which four to put money behind.

How fast is the first batch?

We start the day your brief lands — there is no queue and no kickoff call to schedule. Three business days for the $50 batch, five for Test and Growth, ten for Partner.

What if I hate the first batch?

It cost you $50, so not much is at stake. On any package: email us within 30 days and we refund it in full, and you keep the files either way.

Can you match my brand guidelines?

Yes. Send them at intake. Fonts, colors, tone, claims you can’t make, words you can’t say.

Do you post it for us?

We hand you a calendar and files scheduled against it, and we’ll load them into your scheduler if you give us access. We don’t take over your accounts and we don’t manage your ad spend — your buyer or your in-house team keeps that. If you want someone to run the media too, say so and we’ll tell you honestly whether we’re it.

With AI. The actors are AI characters and the footage isn’t filmed — that’s how the price sits under one creator’s day rate. It isn’t what you’re buying: you’re buying the angles, the formats, the cadence, and a human who decides what dies this week. We label AI content per Meta and TikTok’s requirements and we don’t produce fake testimonials or put words in the mouth of anyone claiming to be a real customer. That line isn’t negotiable.

Some AI video does, and most of the cheap stuff does. We spend most of our production effort on the things that give it away — the speech cadence above all, then hands, faces, text on screen, the grade. Watch the work above and decide for yourself. That grid is the honest answer to this question.

Why not just buy Arcads or Creatify and do this myself?

You should, if you have someone in-house who will decide the angles, write the scripts from your reviews, sit through the renders, cut the tells, and read the results every week to decide what dies. The tool is about $110 a month and we’re not going to pretend otherwise. You’re paying us for the person and the plan, not the model.

Meta generates video for free inside Ads Manager. Why pay you anything?

Because free generation gets you variations of what you already have, and it has no opinion about what to make next. It’s a good reason not to pay anyone for raw volume — which is exactly why we don’t sell it.

Can you promise me a viral hit?

No, and anyone who does is lying to you. Nobody can name the video that pops. What you can do is aim well and take enough shots — angles pulled from what your buyers actually say, formats that already work in your category, and a cadence that gives the algorithm something to chew on. Then read what travelled and make more of it. That’s the whole method, and it’s the closest honest thing to a strategy for going viral.

Do I need to send product samples?

No. We work from your product photography and website. Nothing ships anywhere.

What if my product is regulated?

Tell us at intake. We work in supplements, skincare, and finance, and we’ll write to your compliance constraints. We won’t make claims your legal team hasn’t cleared — and in a few of those categories we’ll tell you to hire a real creator instead.

Do I have to commit to anything?

No. Every package is a one-off — you buy a batch, we deliver it, and that is the end of it unless you buy another. No subscription, no notice period, no card kept on file, nothing to cancel.

Notes

Meta's AI Labeling Policy Explained: When the 'AI Info' Tag Appears Next to Sponsored

Published 21 August 2026

Meta treats AI-generated photorealistic people differently from AI-generated everything else. If your ad has a synthetic human in it, the disclosure is prominent by design.

Two tiers of label, and only one of them is visible

Meta's approach to generative-AI transparency in ads distinguishes between minor AI assistance and AI-generated photorealistic humans. Per Meta's own announcement, ads containing photorealistic AI-generated people carry an "AI info" indicator displayed alongside the Sponsored tag — not hidden in the post's overflow menu where most metadata lives.

That placement matters more than the existence of the label. A disclosure behind a menu is functionally invisible; a disclosure next to Sponsored is part of the creative. You should assume every viewer sees it and write the ad accordingly.

What triggers it

The trigger is a photorealistic person who was generated rather than filmed. Stylised or obviously illustrated characters sit in a different bucket, as do AI-assisted edits to real footage — colour, background, upscaling — which is the category most brands are already using without thinking of it as AI.

Meta also detects generation from industry-standard provenance signals embedded by the tools themselves, so this is not purely self-declared. Meta's own help centre documentation covers the mechanics.

Why this is fine, and how to make it fine

The instinct is to treat the label as a penalty on performance. In practice it's a constraint on *format*, not on results: what suffers is the ad that depended on the viewer believing a real customer was speaking. An ad that reads as an ad, with a presenter, loses nothing by being labelled as produced with AI.

Which is the argument for building the creative around a spokesperson from the start rather than around a fake customer — the label stops being a problem the moment the ad wasn't pretending. Why that model is the safer one, and why our terms commit to it.

Operational notes

Assume the label appears. Don't build a concept that only works if it doesn't.
Don't attempt to strip provenance metadata. It's detectable, it's a policy violation, and the downside is your ad account.
Keep the AI disclosure consistent between the ad and any organic version of the same asset.
Policies here move. Re-check before a big launch rather than trusting a blog post — including this one.

This isn't legal advice. It's an operator's account of rules we work under every day. If you're making a call with real money or real exposure behind it, get it reviewed by someone qualified.