Opten: AI Image & Video Prompt Optimizer - Optimize AI prompts for Midjourney, Seedance & 60+ models

Hey Product Hunt 👋 I'm Vlad, solo founder. I built Opten because I was burning through AI credits like crazy. Here's what kept happening: I'd write a prompt that nailed it in Midjourney, paste the same thing into Kling for a video — garbage. Try Seedance — also garbage, but in a different way. Turns out Seedance ignores negative prompts entirely. Kling needs explicit camera motion. Midjourney prefers short. Nano Banana wants natural-language layout descriptions. Veo 3.1 needs explicit shot framing. Each model is its own language. There are plenty of "AI prompt optimizers" out there, but they all give generic advice — "be more specific," "add lighting." Useless. The real answer is model-specific: what works for one breaks another. So I spent months writing skill files for each model — 61 of them now, covering Midjourney, Nano Banana, Veo 3.1, Kling, Seedance, Flux, GPT Image, Runway, Wan, LTX-2 and 50+ more. Opten reads which model you're on, scores your prompt 0–100 against that model's actual rules, and rewrites it in one click. Works directly inside syntx.ai, higgsfield.ai, freepik.com and magnific.com — no copy-paste, no tab switching. Three things I'd love your feedback on: 1. Which model are you using most right now? Tells me what skill to improve next. 2. What's the worst "generic prompt advice" you've ever gotten? I'll write a counter-rule. 3. Which platform should I support next? (Aitrix, OpenRouter, Flora are on the shortlist) Happy to answer anything — Claude Haiku 4.5 under the hood, Vercel proxy, all skill files are markdown. Thanks for trying it 🚀