Skip to content

Where the effort went

The boring half of the job, done for you

Every one of these starts from material you already have. That is the whole idea.

01

Your script, as a playable board

Hand it the document your dialogue already lives in and you get the board back with the dialogue still in it — every line in its own node, attributed to the right speaker, choices branching where you wrote them. Not a summary and not a rewrite: your words, moved. A finished scene becomes something you can walk through in the time it takes to read this.

02

Your character sheet, as a real entity

Hand it Ryn's character sheet and say "make a Character entity from this". If the project has no Character definition yet, it writes one first — the fields your sheet actually has, typed — and then fills it in with Ryn. The container and the contents both come off your page. Nothing invented, nothing paraphrased. That is fifteen minutes of copying between a document and a form, deleted, and you never had to build the form.

03

Ask your own story

"What happens to Ryn in Act 2?" is a question you can simply ask, because the assistant has read the project. Useful precisely when you are planning Act 3 and cannot hold all of Act 2 in your head at once.

04

Placeholder localization

Not a replacement for your translators, and not pretending to be. But when you need to playtest the German build on Tuesday and the real translations land next month, you can generate a stand-in and get on with testing.

05

Placeholder art

Portraits and images to prototype against, so a board under construction reads like a game instead of a wall of grey boxes. Swap them for the real thing whenever it arrives.

Being straight about it

And yes — it can write

Ask it to generate a board with an original story in it and it will. Because it retrieves from your project first, what comes back tends to fit the cast and continuity you already have rather than arriving from nowhere. Some tools in this space promise they would never do such a thing. Ours would, and you would find that out in five minutes, so there is no point in us saying otherwise.

Whether you work that way is your business. We are not going to ship a lock and call it a principle. We simply did not aim the effort there — a blank-page generator is the one thing this market is not short of, and it is not what was slowing you down.

Regardless of how you use it

What is true of every generation

Reviewed before it lands
Every generation in Thalia's generator tabs stops at a review step. You read it, edit it, and decide whether it enters the project. The assistant works differently — it applies as it goes, and one click reverts a whole conversation.
Shaped like your schema
Ask for a character and you get one shaped like YOUR Character definition — your fields, your types. It generates into the structure you defined rather than around it.
Context when you want it
Searching your project for lore is opt-in per generation. Your project's names always travel with a request, so the output uses them; turn search on, and it goes looking for the rest itself.
Priced in front of you
Estimated cost before you run anything, a per-generation log, and a credit balance in view. No feature spends on its own initiative.
Entirely optional
The canvas, the world model, scripting, play mode, and the engine sync do not depend on a single credit. Ignore all of it and the tool is complete.
Not training data
Your projects are never used as training data.

Why this way

The bottleneck was never the prose

Branching narrative has an unusually bad ratio of craft to clerical work. For every hour spent deciding what a character would say, there is another spent copying that decision into nodes, wiring conditions, keeping a variable name consistent across forty branches, and re-entering a character sheet you already wrote into a form that wants the same fields in a different order.

That second hour is the one people actually complain about, and it is the one nobody had automated — because it is unglamorous, and because doing it well means understanding the writer's own structure rather than generating a fresh one.

Writers do not run out of ideas. They run out of the hours those ideas need.
So that is the hour we went after.