From prompt to clip in 30s — I built SoVideo for people searching “new sora 2”

Hey IH! I’m shipping SoVideo, a template-first AI video tool for makers who keep searching “new sora 2” and just want a first clip in under 30 seconds — no GPU, no long tutorials.

Why I’m building this
Too many tools optimize for demos, not for the moment a creator needs to test an idea today. I care about speed to first value and being honest about limits.

What actually works so far

• Template-first flow → pick a template, paste a prompt, render fast, iterate.

• Lightweight onboarding → get to a preview before any heavy setup.

• Public benchmarks & weekly changelog → ship small, show numbers.

Where it breaks (and what I’m fixing next)

• Fast camera motion & multi-scene continuity still degrade; I’m adding guardrails and clearer prompt tips.

• Long renders on very detailed prompts; working on smarter defaults and queueing.

Roadmap (next 2–4 weeks)

• Batch rendering with sane caps

• “Explain this failure” button with auto suggestions

• A tiny gallery of reproducible examples (copy → run → same result)

If you make short how-to, demo, or social clips, I’d love feedback on the first-run experience. What would remove the next 2 clicks for you?

→ Try it once (and tell me what breaks): https://sovideo.app/?utm_source=IH&utm_medium=post&utm_campaign=launch