Apply for beta →
Menu
How it worksPricingAbout Apply for beta →
How it works

From script to finished video, without losing the thread.

Six steps across three stages. Setup locks characters, places and style once. Scenes turns each narration beat into a checked shot and animates the picture you approve. Film lets you judge the whole video as a viewer would.

  1. Script

    Bring the script

    Paste a finished narration script. VIA splits it into story sections and scenes, one scene per narration beat, and times each one to the voice-over.

    You
    Paste the script
    VIA
    Scenes, sections, timing
    SCRIPT · 83 scenes · 14 sections

    0:00Hey guys, welcome back to Ancient Rome. I'm Marcus, and I have twenty-four hours to survive here on one Roman worker's wage.

    0:09I normally work in finance in the twenty-first century, but earlier today I was dropped into Rome…

    0:21Somehow, I actually found one.

  2. Setup

    Approve the production bible

    VIA proposes the look, a reference sheet for every recurring character (front, side, back, fixed wardrobe) and every location with the features that must never change. Everything later is drawn from these.

    You
    Approve or redo each one
    VIA
    Characters, places, style
    Setup: character reference sheets and a locked Roman street with its fixed features
  3. Scenes

    Direct each scene

    For every beat VIA writes the shot: who is in frame, where, doing what, holding what, from which camera. An independent check compares it with the narration before anything is drawn.

    You
    Read and adjust if you want
    VIA
    Shot written and checked
    A shot description written from the narration
  4. Scenes

    Pick the picture, then animate it

    Stills are drawn from the approved references and judged for identity, place, props and framing. The one you pick becomes the first frame of the clip, so the video shows the same people in the same place.

    You
    Pick the still
    VIA
    Drawing, judging, animating
    Scenes: the chosen still with its other candidates below
  5. Film

    Watch it as a film

    The whole video plays with narration on a scene timeline. Judge pacing and continuity the way a viewer would. Reject a scene and only that scene is redone.

    You
    Watch, reject what is off
    VIA
    Assembly, retakes
    Film: the assembled video playing above its 83-scene timeline
  6. Publish

    Export and publish

    Export the finished video for YouTube. Only approved, current clips make the cut, so nothing outdated slips into the final file.

    You
    Upload it
    VIA
    The final file
    Thumbnail frame of the finished video
    24 Hours in Ancient RomeReady
    9:22 · 83 scenes · narration mixedMP4

You make the creative calls. VIA does the production.

You decide
  • The script and the look
  • Which characters and places are right
  • Which take is right
  • Whether the film works
VIA does
  • Scene breakdown and shot writing
  • Every image, motion and voice prompt, and the model for each shot
  • Checking each result against the script and your references
  • Versions, retakes and assembly

Built for narrated long-form with recurring characters and places. Not for talking-head vlogs, editing live footage or real people's likenesses. See it across a whole production in the 24 Hours in Ancient Rome test.

Make AI video people actually want to watch.

VIA is admitting a small number of creators into the private beta. Tell us what you want to make.

Apply for the beta → How it works