film-theory

WorldWeaver: world state registers for AI storyboards

Story2Board Team··13 min read
WorldWeaver: world state registers for AI storyboards

WorldWeaver is useful for AI storyboarding because it says the quiet part out loud: a generated frame is not a world state. If two agents stand in the same Minecraft world, each first-person video stream only sees part of the truth. The model still needs one shared memory of where both agents are, what changed, and what should carry into the next generated chunk.

Mo et al. (2026) call that memory a world state register. The idea is technical, but the product lesson is plain. If a storyboard system wants consistent characters, locations, props, and camera logic, it needs a state layer that lives outside the pixels.

Two agent video streams writing into shared world state registers before generating consistent frames.

What the paper does

WorldWeaver is a streaming multi-agent autoregressive diffusion model. It starts from the Solaris multi-agent Minecraft setting: two players, synchronized first-person video streams, action inputs, and a shared simulated world.

The baseline approach carries forward recent frames through a local KV cache. That helps temporal continuity, but it doesn't give the model a durable shared state. A frame cache is biased toward what was recently visible. It may not remember what the other agent saw, what happened off-camera, or what the global layout says should be true.

WorldWeaver adds a set of learnable register tokens to the rollout. After each generated chunk, the model updates those registers. The next chunk reads both the frame cache and the latest register state.

Comparison between frame-only streaming video diffusion and WorldWeaver's register-based rollout.

The registers are trained to hold three kinds of world information:

Register supervisionWhat it storesWhy it matters
Agent statsposition, velocity, orientationkeeps individual motion and facing direction verifiable
Bird's-eye viewglobal layout and relative placementgives the model an allocentric map, not just egocentric views
Scene textlanguage description of the current statekeeps high-level semantic changes readable

The prediction heads are only used during training. At inference time, the model keeps the registers but discards the supervision heads. That matters because the paper is not adding a visible text planner beside the model; it's trying to make the hidden state inside the generator more grounded.

Why frame memory fails

We see the same pattern in storyboards, just with still panels instead of Minecraft videos.

A storyboard panel is an observation. It is not the whole scene. It may show a close-up of a key, but not the kitchen layout. It may show a character's face, but not where the other character is standing. It may show a door closing, but not whether the stolen prop is now gone from the table.

If the next panel is generated from the previous image plus a prompt, the model can imitate surface appearance while missing state. The face may stay close enough, but the room flips. The prop returns. The eyeline points nowhere.

This is why WorldWeaver lands near our existing reading of CANVAS and GEST-Engine. CANVAS argues that characters, locations, and object state need explicit continuity tracking. GEST-Engine argues that video should trace back to event state. WorldWeaver makes a similar argument inside streaming diffusion: the rollout needs a persistent state pathway.

Not a longer prompt. A state pathway.

The rollout in plain English

WorldWeaver trains in three stages.

First, it adapts a video diffusion model to synchronized two-agent clips with bidirectional attention. This teacher sees the full clip and learns the shared scene structure.

Second, it turns the model into a causal streaming student. The student sees only past context, generates the next frame, and updates the world state registers. At this stage the registers receive explicit supervision from agent stats, bird's-eye-view features, and scene text.

Third, it uses self-forcing. The model rolls out on its own generated history rather than only ground-truth frames. That exposes both frame drift and register drift during training. The paper notes that DMD-style self-forcing alone was not enough; register supervision still mattered for stabilizing long-horizon rollout.

The architecture also changes. WorldWeaver uses a Mixture-of-Transformers design so frame tokens and register tokens can use separate weights while still attending together. The paper's explanation is sensible: pixels and state are doing different jobs. Asking the same dense transformer path to render frames and maintain a semantic world register can create conflict.

The numbers

The experiments use about 126 hours of synchronized two-player Minecraft recordings. Evaluation is on the Solaris test split. The paper reports VLM accuracy and FID across five categories: Movement, Grounding, Memory, Building, and Consistency. It also reports an auxiliary WorldScore, which combines VLM accuracy and FID. The authors warn that WorldScore is unbounded and scale-sensitive, so it should be read as a ranking aid rather than a universal metric.

Still, the pattern is clear.

MethodMovement VLMGrounding VLMBuilding VLMConsistency VLMWorldScore
Frame concat77.153.10.049.549.1
Solaris79.781.39.457.881.0
WorldWeaver82.893.828.176.6105.1

The biggest wins are not cosmetic. Grounding rises from 81.3 to 93.8. Building rises from 9.4 to 28.1. Consistency rises from 57.8 to 76.6. Those are the categories where a hidden state should matter.

Bar chart showing WorldScore gains from Solaris to registers-only, scene-text supervision, and all supervision.

The ablations are more interesting than the main table:

VariantWorldScoreWhat changed
Solaris baseline81.0no world state registers
Registers only93.8persistent register slots, no explicit targets
+ Scene text103.2semantic grounding helps most among single signals
+ All supervision105.1agent stats, bird's-eye view, and scene text together

Registers alone help. That suggests the model benefits from having a dedicated state slot even before the state is fully supervised. But the best result comes when the registers are told what to encode.

That is the lesson I would keep.

What maps to Story2Board

Story2Board doesn't need to copy a Minecraft simulator. We don't have a ground-truth bird's-eye camera for a drama scene in a kitchen. We don't get exact velocity vectors for every character. And most storyboard projects start from script text, reference images, and director notes rather than a game engine.

But the mapping is still strong.

Mapping WorldWeaver register signals to Story2Board fields for characters, props, locations, narrative purpose, and reference URLs.

WorldWeaver conceptStory2Board equivalent
Agent statscharacter position, facing, action, eyeline, emotional state
Bird's-eye viewlocation layout, scene geography, camera axis, blocking map
Scene textnarrative purpose, beat summary, dialogue context, sound cue
Register tokeneditable shot state and project memory
Register updaterevise shot fields after each generated panel

This is close to how our shot model already wants to work. A shot is not just an image URL. It has action, shot size, angle, camera movement, equipment, focal length, lighting, duration, emotional tone, narrative purpose, dialogue, sound design, notes, characters, props, and reference URLs.

WorldWeaver gives a research vocabulary for why those fields matter. They are not paperwork. They are the state that pixels cannot reliably preserve by themselves.

A better storyboard memory contract

If we translate the paper into product language, every shot should answer four state questions:

QuestionExample
What must persist?Maya wears the red raincoat; the kitchen window is behind her left shoulder
What changed?the key moved from the table to her hand
Who can see it?Maya sees the key; Jesse is off-screen and does not
Which later shot depends on it?shot 07 must show Jesse reacting before he sees the key

That last question is usually missing. Storyboard tools often track what appears now. They rarely track what a later panel will depend on. World state registers are a reminder that state is useful because it conditions the next chunk.

For an AI storyboard, that means the memory layer should be forward-looking. A prop state is not just a note about panel 03. It is a constraint on panel 04, panel 08, and any return to the same room.

Where the paper is limited

The limitation section is short but important. The authors say their main improvement comes from additional state supervision, and real-world state is much more complex and not directly available.

That's the hard part. Minecraft gives you position, velocity, orientation, action inputs, overhead views, and synchronized streams. A live-action screenplay does not. A creator may write "she feels watched" and never specify where the watcher is. A reference image may imply a room layout but not define it. A prop may matter emotionally before it matters visually.

So the direct research path is not "add registers to a storyboard product." The product path is more modest:

  • keep state fields editable;
  • preserve reference images and previous generations;
  • ask the director when state is ambiguous;
  • separate visible image prompts from hidden continuity notes;
  • evaluate generated panels against the state contract.

Honestly, that is more useful than pretending the model can infer everything.

Why this belongs with CANVAS, GEST, and EvolvingWorld

Several recent papers are converging on the same design direction:

PaperState layer
CANVAScharacter, location, and object continuity tables
GEST-Engineevent graph with actor/object/location/time state
EvolvingWorldopen-schema character and world profiles
WorldWeaverhidden world state registers inside streaming diffusion

They operate at different layers. CANVAS works at the storyboard-image pipeline level. GEST uses simulation and annotation. EvolvingWorld works with literary world state. WorldWeaver works inside a video diffusion rollout.

The shared bet is the same: generated media needs state that survives from one moment to the next.

What we would build from this

For Story2Board, the near-term version is not a neural register. It is a structured register the creator can inspect.

The co-director can maintain:

  • a character state row for each recurring character;
  • a location state row for each reusable space;
  • a prop state row for important objects;
  • a shot state row for action, camera, sound, and narrative purpose;
  • a reference row for images that should condition future panels.

Then each generated panel can be reviewed against that state. Did the character drift? Did the room flip? Did the prop reset? Did the camera break the dialogue axis?

That is less glamorous than a hidden register bank. It is also shippable.

And later, if storyboard models become more like streaming world models, those explicit fields become training targets. That is the bridge WorldWeaver points toward: a storyboard tool can collect state in human-readable form today, and future generators can learn to carry it as internal state tomorrow.

Bottom line

WorldWeaver is not a storyboard paper. It is a multi-agent video diffusion paper in Minecraft. But it is directly relevant because it attacks the same failure mode that makes AI storyboards unreliable: the model sees frames, but it forgets the world.

The useful lesson is not only that registers improve a benchmark. It is that consistency improves when the system has a dedicated place to store state and explicit signals that teach that state what to remember.

For AI storyboards, that means every panel should be more than a prompt and an image. It should be a state update.

References

Mo, S., Li, Y., Leng, Z., Singh, K. K., & Zhou, B. (2026). Streaming Multi-Agent Autoregressive Diffusion Model with World State Registers. arXiv preprint arXiv:2607.21594. https://arxiv.org/abs/2607.21594

Savva, G., Michel, O., Lu, D., Waiwitlikhit, S., Meehan, T., Mishra, D., Poddar, S., Lu, J., & Xie, S. (2026). Solaris: Building a Multiplayer Video World Model in Minecraft. arXiv preprint arXiv:2602.22208.

All diagrams were created by the Story2Board team for this analysis.

Related Posts

How Visual Continuity Works in AI Storyboards: Lessons from CANVAS

A practitioner's analysis of the CANVAS paper — how global planning, visual memory, and prop-state tracking solve the biggest challenge in AI-generated storyboards: maintaining consistency across shots.

Storyboard consistency checklist for AI-generated boards

Use this AI storyboard consistency checklist to track characters, locations, props, camera continuity, and review notes across panels.

World Narrative Model: controllable video from physical world blueprints

WNM splits video generation into a world controller and a frozen renderer. We read the paper for Storyboard and explain why that matters for shot planning.

The GEST-Engine: event graphs as a storyboard control layer

The GEST-Engine turns event graphs into annotated multi-actor video. We read the paper for Story2Board and explain why explicit story state matters.

EvolvingWorld: co-evolving character and world state for storyboards

EvolvingWorld models characters and literary worlds as open-schema state. We read the paper for Story2Board and map it to storyboard continuity.

ShotPlan: learnable planning tokens for multi-shot video

ShotPlan controls hard cuts and camera timing with learnable planning tokens. We read the paper for Story2Board's multi-shot preview workflow.

GraphVid: interaction graphs for AI storyboard motion

GraphVid controls image-to-video generation with directed interaction graphs. We map the paper to Story2Board's shot fields, graph edits, and preview workflow.

Ready to create your storyboard?

Turn your ideas into professional storyboards with Story2Board — the intelligent director assistant.

Try Story2Board Free