Describe a video — your plan and preview appear here.
Chat on the left, your plan and preview on the right. Here's a real session — a request, the agent's plan, your edits, the render, and revisions to the finished video.
Not a chatbot that hands you tips — a creative agent that scripts, plans, casts and renders, then refines on your word.
Talk to your studio in plain language — powered by Claude. No timelines, no menus. Describe the video and the agent takes it from there.
Give it a topic, a link or a rough idea and the agent writes the hook, the story and the voiceover — tuned to your goal and audience.
Every video starts as a clear scene-by-scene plan. Each scene shows its duration, what's on screen, and the exact voiceover — edit it before a single frame renders.
Cast an AI avatar, apply your brand colors and fonts, and choose a voice — the agent applies them across every scene in the plan.
Even after the video is made: translate it, rewrite the hook, add a scene, swap the voice or avatar — just ask and a new cut is generated.
Write the plan yourself, or build it with the agent. Generate fresh plans, tweak existing ones, or duplicate a winner — you stay in control.
A plan is a list of scenes the agent writes and you can edit. Each scene carries three things:
How long the scene runs, so the whole video lands on your target length.
0:06A description of what's on screen — the shot the agent generates or assembles.
Phone glows awake; the logo ignitesThe exact line spoken in the scene, in your chosen voice and language.
“Meet Lumi — your night, reimagined.”Generated a video and want it different? Ask in the chat — the agent re-plans only what changed and renders a new cut.
An AI video agent makes videos through conversation: you describe what you want, it plans and generates the scenes across 30+ models, and revises on request. Vivideo's agent is the chat-based way to make a video, no editor needed.
A plain generator makes one clip from one prompt. The agent plans a whole multi-scene video, generates it, and keeps revising by conversation — closer to briefing a teammate than filling a prompt box.
No — you make and refine the entire video by chatting. The agent handles scenes, models, captions and pacing, so there's no timeline to learn.
Yes — just ask. Say "make the intro shorter" or "swap that shot" and the agent revises the relevant parts, as many times as you like.
The agent runs across 30+ models — Sora, Veo, Kling and more — picking the right one per shot, so you get the best look without choosing yourself.
Yes — you can make a video with the AI agent free, with daily generations and no credit card. A small upgrade unlocks watermark-free exports.
Yes — take the agent's video into the editor for fine control any time, or switch to manual mode to drive the models yourself.

Vivideo's AI video agent makes video by conversation — describe what you want, and it plans, generates and assembles the whole thing across 30+ models. The answer to "make a video without operating an editor."
Delegate the video instead of building it, scene by scene.

The agent shows you a scene-by-scene plan before rendering, then revises on request — a new hook, a swapped shot, a shorter cut. Iteration at the speed of conversation, no timeline.
Say what's wrong and the agent re-generates just that part.

Chatting is one way to make an AI video on Vivideo — alongside one-prompt auto-generate and hands-on manual mode. Free to try, with watermark-free exports on an affordable plan.
Pick the workflow that fits — talk, one prompt, or full control.
One place to plan, generate and revise a whole video by chat.

Anyone who'd rather describe a video than build one:
Each utility here is a doorway into the full Vivideo studio — create, edit and localize video in the same place.
Watch a brief become a finished video through chat — then make your own free on Vivideo.