JSON to Video: Generate Veo & Sora videos from structured JSON prompts

I’m the maker of JSON to Video. Over the last year, every AI video tool I tried felt like pulling a slot machine: write a poetic prompt, hit generate, then spend credits hoping the model understands what “cinematic product shot with warm lighting and upbeat music” means. Changing one word could completely change the video. Great for demos, terrible for production. JSON to Video is my attempt to fix that for people who actually have to ship ads, campaigns and client work. Instead of chatting with the model, you define a schema: – subject: what’s on screen – camera: framing, movement, lens – lighting: time of day, mood, contrast – audio: music, ambience, VO – duration + guardrails Then you choose Veo 3.1 or Sora 2 and let the model execute that plan. Same JSON, same look – even when you swap the underlying product or script. Who I built this for: – Performance marketers who need ten variations of a winning concept, not ten random surprises – Agencies and in-house teams who want juniors to work inside safe, on-brand templates – Automation nerds who want to generate videos from spreadsheets, APIs or workflows instead of babysitting a prompt box I’d love feedback from the PH community on three things: What’s missing for this to plug cleanly into your existing workflow (APIs, webhooks, n8n/Make/Zapier blocks, etc.)? What knobs do you actually want in the JSON (camera, pacing, transitions, titles) vs what should stay automatic? Would you prefer a “builder” UI that outputs JSON, or are you comfortable living directly in the schema? I’m actively iterating, so if you’re running real campaigns or client work and want predictable Veo / Sora videos instead of prompt roulette, I’d love to hear how you’d abuse this.