An AI web series maker built around the season

The series problem

Making one good AI episode is achievable. Making the tenth one match the first is where most projects fall apart. Characters get re-described and re-invented, a recurring apartment changes layout, a voice shifts. Viewers notice, even if they can't say what changed.

In ShowMaker a series is a container. Its cast, wardrobe, locations, props and voices belong to the series, and every episode draws on them. You don't rebuild the lead for each episode. You reference her.

The same goes for places. A recurring apartment, office or restaurant is a location with reference stills, so the audience sees the same room every time the story returns to it. Props and wardrobe work the same way: the ring from episode one is still the same ring in episode twenty.

Planning a season before you shoot

Drama Builder starts from tropes, the recurring patterns audiences of serialized drama respond to, and builds a premise. From there it builds a series skeleton and an episode outline, and once you're happy you lock in the series. It writes real screenplays for the episodes, which land in the Script editor.

Writers' Room rates the opening hook and improves it in rounds, because in a series the first episode has to sell all the others.

The outline is where a season's shape gets decided: where the big reveals land, which episodes end on the hardest cliffhangers, which characters arrive when. Changing that on paper costs nothing. Changing it after ten episodes are rendered means re-shooting.

• Premise and tropes
• Series skeleton and episode outline
• Screenplays for each episode
• Shared cast and assets across the season

Producing at episode pace

Each episode goes through the same path: Scene Flow for script and shots side by side, stills per shot, voices, the free narrated animatic, then video for the shots that need it. Because the assets are shared, later episodes go faster, since the cast is already built.

Use the free narrated animatic as the gate for each episode. If an episode drags or its cliffhanger is soft, fix it while it's still stills and voices. Once an episode works as an animatic, render only the shots that need motion and leave the rest as still-motion where it serves.

Choose the video model per shot to keep the season affordable. A full 67-second vertical episode with a consistent cast and 8 multishot panels cost under $7 in credits. Most of the cost comes from re-rendering shots, so approving the script, stills and animatic first matters more than any per-clip price.

Vertical or horizontal

Short serialized drama is usually vertical, 9:16, for phone viewing. Longer web series often run 16:9. A Hybrid project makes the same story in both, each framed for its own shape, so you aren't cropping one into the other.

Decide this early. Framing is decided when the picture is made, so you can't just change it at export. A close-up composed for a tall frame and one composed for a wide frame are different pictures. If you might want both, set the project up as a Hybrid from the start.

Either way, keep episode length consistent within a season. Viewers settle into a rhythm, and an episode that runs twice as long as the others feels like a different show. A tight outline is what keeps each episode to its length.