Built a thing. Want people who think about interactive fiction to poke at it.
How it works - you type what your character does, a narrative engine resolves it against the scenario’s world rules and characters. It generates what happens next as prose and a comic panel sequence. No preset choice menus. The story branches from your actual input.
Link to The Krale Road scenario in Fractal
The scenario I am linking is The Krale Road in the horror/mystery genre, but you are welcome to explore the other scenarios from there (just click the ← SCENARIOS link at the top left). The premise: Your car died on a back road in Lousiana late at night. Someone from your group knocked on a farmhouse door and hasn’t come back.
It’s rough in places, and some narrative branches resolve better than others. If you feel like your actions aren’t being executed fully or truly moving the narrative in a certain direction, then it might be the specific scenario rules that are too constricting. This particular one has a three-act structure that might be too restrictive.
A few things to know going in:
- Turns take 1-2 minutes to generate (narrative first, then image). I plan to add notifications so you can sort of fire and forget until generation is done.
- The narrative engine and image generation use AI. The scenarios themselves were built with AI assistance, which is kind of necessary for them to work well with the engine.
- Sign up is required to play and requires invite codes
Each code gets you 25 turns (they are one-time use so please only claim one):
INTF329
INTF213
INTF393
INTF023
INTF652 claimed
If all codes are used up, let me know and I’ll make more.
Your generated branches can be published. That’s partly why you might see existing panel sequences in the scenario before you’ve played it. Other players’ published brances are visible in the branch tree.
If you’re willing, publish your panels rather than keeping them private. The branch tree gets more interesting the more paths exist in it, and right now it’s pretty sparse.
This is roughly what it looks like:
One thing I’m genuinely curious about: does the visual component add anything for you, or does it get in the way? The images are AI-generated and there is some work still needed for character, environment, etc consistency (more of a technical problem to solve, but not tackling it in alpha yet).
Also thinking about building a scenario authoring tool so people can create their own. The scenario format is structured (world ruules, character mechanics, phase logic, endings). Curious if that’s something people here would actually use, and what you’d want to be able to control.