What is decodx?
decodx turns a committed, human-authored spec into a finished video — final.mp4 plus captions, chapters, and a thumbnail — deterministically. You author scenes the way you author source code, and re-render freely because expensive assets are cached by content hash.
What counts as an "AI minute"?
An AI minute is a minute of newly generated assets — narration, screen or browser captures, and avatar clips. It is what your plan meters. Re-rendering after a spec tweak reuses cached assets and never spends AI minutes, so iterating is free.
Why do avatar minutes count more?
Talking-head (avatar) clips cost several times more to generate than plain narration, so one avatar minute counts as four AI minutes. That keeps a single, legible meter instead of separate credit pools.
Do re-renders cost anything?
No. Every finished asset is keyed by a content hash, so changing one scene re-renders only what changed and pulls everything else from cache. Re-renders are free on every plan, including Free.
Is there a watermark?
The Free plan adds a small decodx watermark to renders. Every paid plan — Creator, Team, and Enterprise — removes it. Upgrading re-runs only the final compose step, so your video is watermark-free in seconds without re-buying any assets.
How does billing work — monthly or annual?
Both. Annual billing saves about 20% versus paying monthly. Plans are metered per organization, and members are pooled across your whole workspace.
Can my AI agent create videos on its own?
Yes — decodx is built for agents. Every command is non-interactive and every artifact is schema-validated, so an agent can author a spec, render it, and read back structured results. See "Video skills for agents" for how it works.
What happens when I hit my plan limit?
On paid plans a render is blocked with a clear upgrade prompt once you reach your monthly AI-minute limit — no surprise charges. Concurrency and member limits work the same way: they cap what runs at once rather than failing silently.