Morpheus WireStory Atom
Understand any story as a map.
Story Atom reads a script, an outline or a one-line idea and turns it into a living map of the story — who is in it, where it happens, what changes, and how it should look and sound.
01 · What it is
Your story, atomised.
Story Atom dissects unstructured story into a structured, time-anchored graph — so an AI can work on your story with full knowledge of it.
Story Atom02 · The six levels
Six levels of DNA.
Four irreducible roots - Time, Space, Genre, Theme - sit at the top. Everything else branches from them, level by level, down to Expression: how the story is finally seen and heard. 35 atom types, each built on what came before.
03 · Getting a story in
Time is measured by reading.
The source is cut every 25 tokens — ignoring lines, sentences and scenes. Each slice becomes a point in time: T1, T2, T3…
25 tokens · 25 tokens · 25 tokens · …
04 · Decisions, not writing
Thousands of tiny yes-or-no calls.
Every slice is checked on its own, in parallel. A decision model doesn't write — it only decides. Smaller and far faster than an LLM.
LLM still deciding…
Time to judge the grid
Laya by Convai Innovations. Speeds as reported in early tests; animation not to scale.
05 · Identity
Same name. Different person.
An investigator reads a wider window and answers one narrow question: is the Mr. Waters at T12 the same person as Ben? The clues say no.
INVESTIGATOR WINDOW — much wider than one slice
T7 “Mr. Waters, the usual?” Ben nods and takes the glass.
T12 “Ben — your father called. Mr. Waters wants you home.”
Ben Waters
aka Ben · Mr. Waters?
06 · Time points become beats
Neighbours vote. Beats form.
Each pair of neighbouring time points is scored level by level. Strong evidence keeps them together; when the core changes, the beat ends.
T1 vs T2 · score × weight = points
07 · The result
One node per real thing.
Every person, place and object appears once — pointing at every moment it's in. Beats span the timeline. Behind every point is the original text.
08 · Why the graph
Ask anything. Get a cited answer.
Without the graph, an AI rereads the whole script and guesses. With it, the agent touches one time point and gets the verbatim source around it - a 1,000-token touch diameter, overlapping its neighbours so nothing falls between the cuts.
Without Story Atom
Rereads ~30,000 tokens, every question
Story Atom reads only this: ~1,000 of 30,000 tokens
"Probably the bar?" · no source
With Story Atom
Ben nods and takes the glass. Julie slips out to the street.
The door opens - "Ben, your father called. Mr. Waters wants you home."
Ben looks to the door. The street outside is empty…
"On the street." · cites T8, from the prior tail
Verbatim source, never a summary · ~1,000 tokens per touch, never the whole script · any agent can touch the graph and pick up full continuity.
09 · What the graph powers
One graph. Every room reads it.
Story Atom does one thing: it maps the DNA of your story into a graph an agent can understand, grow and reshape. Each Storyhouse tool reads that same graph in its own way. Tap to explore.
Lucid LinesDocument · Scene 1
Ben pushes through the door of the bar, Ben
scanning for Julie. Julie
She's at the counter, glass in hand. Glass
Julie slips out to the street. Street
Continuity: Julie is at the bar and on the street at T5.
The graph grows as you write. Explore your own characters, plots and threads at a glance.
Change the graph and this regenerates - re-editable, redeployable.
10 · The engine
Every room reads from it.
Story Atom isn't a product on its own. It's the DNA every Storyhouse room reads from — and writes back to.
01 · What it is
Your story, atomised.
Story Atom dissects unstructured story into a structured, time-anchored graph — so an AI can work on your story with full knowledge of it.
Story Atom02 · The six levels
Six levels of DNA.
Four irreducible roots - Time, Space, Genre, Theme - sit at the top. Everything else branches from them, level by level, down to Expression: how the story is finally seen and heard. 35 atom types, each built on what came before.
03 · Getting a story in
Time is measured by reading.
The source is cut every 25 tokens — ignoring lines, sentences and scenes. Each slice becomes a point in time: T1, T2, T3…
25 tokens · 25 tokens · 25 tokens · …
04 · Decisions, not writing
Thousands of tiny yes-or-no calls.
Every slice is checked on its own, in parallel. A decision model doesn't write — it only decides. Smaller and far faster than an LLM.
LLM still deciding…
Time to judge the grid
Laya by Convai Innovations. Speeds as reported in early tests; animation not to scale.
05 · Identity
Same name. Different person.
An investigator reads a wider window and answers one narrow question: is the Mr. Waters at T12 the same person as Ben? The clues say no.
INVESTIGATOR WINDOW — much wider than one slice
T7 “Mr. Waters, the usual?” Ben nods and takes the glass.
T12 “Ben — your father called. Mr. Waters wants you home.”
Ben Waters
aka Ben · Mr. Waters?
06 · Time points become beats
Neighbours vote. Beats form.
Each pair of neighbouring time points is scored level by level. Strong evidence keeps them together; when the core changes, the beat ends.
T1 vs T2 · score × weight = points
07 · The result
One node per real thing.
Every person, place and object appears once — pointing at every moment it's in. Beats span the timeline. Behind every point is the original text.
08 · Why the graph
Ask anything. Get a cited answer.
Without the graph, an AI rereads the whole script and guesses. With it, the agent touches one time point and gets the verbatim source around it - a 1,000-token touch diameter, overlapping its neighbours so nothing falls between the cuts.
Without Story Atom
Rereads ~30,000 tokens, every question
Story Atom reads only this: ~1,000 of 30,000 tokens
"Probably the bar?" · no source
With Story Atom
Ben nods and takes the glass. Julie slips out to the street.
The door opens - "Ben, your father called. Mr. Waters wants you home."
Ben looks to the door. The street outside is empty…
"On the street." · cites T8, from the prior tail
Verbatim source, never a summary · ~1,000 tokens per touch, never the whole script · any agent can touch the graph and pick up full continuity.
09 · What the graph powers
One graph. Every room reads it.
Story Atom does one thing: it maps the DNA of your story into a graph an agent can understand, grow and reshape. Each Storyhouse tool reads that same graph in its own way. Tap to explore.
Lucid LinesDocument · Scene 1
Ben pushes through the door of the bar, Ben
scanning for Julie. Julie
She's at the counter, glass in hand. Glass
Julie slips out to the street. Street
Continuity: Julie is at the bar and on the street at T5.
The graph grows as you write. Explore your own characters, plots and threads at a glance.
Change the graph and this regenerates - re-editable, redeployable.
10 · The engine
Every room reads from it.
Story Atom isn't a product on its own. It's the DNA every Storyhouse room reads from — and writes back to.
01 · What it is
Your story, atomised.
Story Atom dissects unstructured story into a structured, time-anchored graph — so an AI can work on your story with full knowledge of it.
Story Atom02 · The six levels
Six levels of DNA.
Four irreducible roots - Time, Space, Genre, Theme - sit at the top. Everything else branches from them, level by level, down to Expression: how the story is finally seen and heard. 35 atom types, each built on what came before.
03 · Getting a story in
Time is measured by reading.
The source is cut every 25 tokens — ignoring lines, sentences and scenes. Each slice becomes a point in time: T1, T2, T3…
25 tokens · 25 tokens · 25 tokens · …
04 · Decisions, not writing
Thousands of tiny yes-or-no calls.
Every slice is checked on its own, in parallel. A decision model doesn't write — it only decides. Smaller and far faster than an LLM.
LLM still deciding…
Time to judge the grid
Laya by Convai Innovations. Speeds as reported in early tests; animation not to scale.
05 · Identity
Same name. Different person.
An investigator reads a wider window and answers one narrow question: is the Mr. Waters at T12 the same person as Ben? The clues say no.
INVESTIGATOR WINDOW — much wider than one slice
T7 “Mr. Waters, the usual?” Ben nods and takes the glass.
T12 “Ben — your father called. Mr. Waters wants you home.”
Ben Waters
aka Ben · Mr. Waters?
06 · Time points become beats
Neighbours vote. Beats form.
Each pair of neighbouring time points is scored level by level. Strong evidence keeps them together; when the core changes, the beat ends.
T1 vs T2 · score × weight = points
07 · The result
One node per real thing.
Every person, place and object appears once — pointing at every moment it's in. Beats span the timeline. Behind every point is the original text.
08 · Why the graph
Ask anything. Get a cited answer.
Without the graph, an AI rereads the whole script and guesses. With it, the agent touches one time point and gets the verbatim source around it - a 1,000-token touch diameter, overlapping its neighbours so nothing falls between the cuts.
Without Story Atom
Rereads ~30,000 tokens, every question
Story Atom reads only this: ~1,000 of 30,000 tokens
"Probably the bar?" · no source
With Story Atom
Ben nods and takes the glass. Julie slips out to the street.
The door opens - "Ben, your father called. Mr. Waters wants you home."
Ben looks to the door. The street outside is empty…
"On the street." · cites T8, from the prior tail
Verbatim source, never a summary · ~1,000 tokens per touch, never the whole script · any agent can touch the graph and pick up full continuity.
09 · What the graph powers
One graph. Every room reads it.
Story Atom does one thing: it maps the DNA of your story into a graph an agent can understand, grow and reshape. Each Storyhouse tool reads that same graph in its own way. Tap to explore.
Lucid LinesDocument · Scene 1
Ben pushes through the door of the bar, Ben
scanning for Julie. Julie
She's at the counter, glass in hand. Glass
Julie slips out to the street. Street
Continuity: Julie is at the bar and on the street at T5.
The graph grows as you write. Explore your own characters, plots and threads at a glance.
Change the graph and this regenerates - re-editable, redeployable.
10 · The engine
Every room reads from it.
Story Atom isn't a product on its own. It's the DNA every Storyhouse room reads from — and writes back to.
Where it fits
From story to finished film.
Next room
SpatialLang
