Why I Built DebaterX
I wanted mascots to fight. The tooling didn't exist. So here we are.
I've been watching brand rivalries play out in ads for as long as I've been paying attention to advertising. Burger King vs. McDonald's. Apple vs. Microsoft. Mac vs. PC. The best ads were almost always two-brand arguments — never one-brand monologues. And yet the vast majority of brand content was one-brand monologues, because putting two brands in one room required lawyers, licensing, and a six-month negotiation cycle.
So for years, the best format in advertising was structurally unavailable to most brands. That frustrated me. It's why DebaterX exists.
What changed
Two things. First, the mascot economy got both more valuable and more volatile. Brands noticed that TikTok was rewarding character-driven content, so they dusted off old mascots or invented new ones. Duolingo's owl. Wendy's. The Planters revival. Lots of movement in mascot land.
Second, AI video caught up to the point where putting two recognizable characters in the same frame — with passable fidelity — became possible in minutes rather than months. That's a new capability. It hadn't existed six months before I started building.
The intersection of those two trends is the opportunity. Mascots matter more than they did. AI can put them in scenes more cheaply than any other medium in history. The infrastructure was missing. I built it.
What it does
DebaterX takes two brand inputs — either mascots with existing imagery or brands that need new characters — and generates short-form video debates between them. The output is 9:16 video, ready to post, 25-60 seconds.
The workflow is simple by design. Pick two brands. Pick a debate topic or let the system suggest one. Tune tone and audience. Generate. Ship.
Under the hood it's doing a lot. Dialogue generation with character rules. Image generation with reference anchoring. Video synthesis with consistency controls. Audio with TTS voice casting. Mux-backed delivery. All the infrastructure a normal agency would need three vendors to stitch together.
Who it's for
Small and mid-sized brands with mascots that aren't getting enough reps. Agencies who want to spec ideas before pitching. Creators who work in brand-parody territory. Product teams who want to test marketing angles cheaply.
Not for: enterprise brands with locked-down legal processes. Those brands already have agencies and budgets. DebaterX's advantage is speed, and enterprise can't absorb speed.
What I'm still figuring out
How to price it. How to position it against agency workflows. How to handle mascot likeness rights in edge cases. How to build the component library of premises so the product gets better each quarter.
Every week I learn something that forces me to rewrite a section of the product. That's the correct state for a startup at this stage. If I had everything figured out, I'd be building something too small.
What I'm sure about
Brand debate content is underbuilt as a format. AI video makes it buildable. Most brands don't know they want this yet. When they find out, the default tooling should be DebaterX.
That's the bet. That's why I built it.