Solo tabletop play runs on oracles: you ask the dice a question, the table answers, and you narrate the consequence. It produces stories with genuine surprise in them — and then the session ends and the story lives in a notebook nobody else can play. Converting a solo run into a branching adventure is the act of turning your oracle’s answers into authored choices: the roll you got becomes one path, the rolls you didn’t become the others.
StorySplice gives that conversion a workspace. Encounters, rooms, and journal entries become scenes on a visual map; the oracle’s possibility space becomes choices you write deliberately instead of rolling blindly. The result is a D&D-style solo adventure — or a journaling game, or a dungeon crawl — that anyone can play in a browser, no system knowledge, no dice, no account.
A random table is a branch point that hasn’t been written yet. Where solo play asks "does the sentry notice me? (roll)", an authored adventure asks the player "distract the sentry, or wait for the watch change?" — same fork, but now each arm carries intention. Converting well means being selective: not every roll from your session deserves a branch. Keep the forks where the fiction genuinely divided, collapse the rest into prose, and spend your branching budget on decisions a player will feel. The structure guide is effectively a solo GM’s prep manual for this.
The Hollow Crown. A disgraced knight re-enters the barrow where her order was broken, hunting the regalia that would restore her name. Classic solo dungeon crawl shape: the barrow is a map of scenes, wounds are tracked by which passages remain open to you, and three of the seven endings leave the crown exactly where it was found.
Ninety Days of Winter. A journaling survival game recast as branching prompts. Each entry is a day at the frozen station; each entry ends with a decision — ration, repair, or radio — that selects the next prompt. The player writes nothing but chooses everything, and the station’s fate accretes from ninety small refusals.
The map is the dungeon key and the flowchart at once — drag encounters around, zoom out to see the whole delve, and let validation catch the corridor that connects to nothing. Expand with AI is unusually apt here: point it at a dead-end choice ("pry open the sealed reliquary") and it drafts a 2–5 scene chain you then rework, which is very close to asking an oracle and editing its answer. Splinter can take a pasted session journal and propose where it could have gone differently — your actual play becomes the spine, the counterfactuals become branches.
Honesty for the mechanically minded: StorySplice has no dice, no stats, no character sheets, and no random tables at play time. The published adventure is deterministic — every branch is authored. If your design needs live randomness or tracked attributes, write the structure here and export to Twee (Pro) for scripting in Twine. For purely choice-driven adventures, which most converted solo runs turn out to be, nothing is missing.
No — play is purely choice-based, with no randomness or tracked attributes. Design the consequences into the branches instead, or export to Twee (Pro) and add mechanics in Twine.
Yes, and Splinter is built for it: paste the linear session log and the AI proposes 4-7 points where events could have branched, then scaffolds the structure for you to rewrite.
No. The published adventure is self-contained prose and choices, playable in any browser without accounts, dice, or rulebooks — a good way to let friends experience a system’s fiction.
Yes. Twine 2 HTML and Twee files import directly; passages become scenes and links become choices. Macros and scripting are stripped, so choice-driven adventures convert cleanest.
Creating and publishing are free with 10 AI calls a day. Pro ($6/mo or $48/yr) adds 200 daily calls, model choice, Twine/Twee export, and reader analytics.