loomlarge

01 / 02

First hello

A calm first hello.

Loom Large opens with presence instead of chrome: an avatar on the left, a polished glass story rail on the right, and just enough explanation to make the social layer feel immediate.

Why this works

Lead with the face so attention feels mutual, not technical.
Use the right rail as a guided story instead of a dense feature dump.
Keep the tone premium, calm, and obviously responsive.

Neck-up control

Tune the social response.

Advance one step and the title settles smaller while the rail turns practical: eye follow and head follow live together so visitors can test a single expressive behavior without opening the full workbench.

Eye follow 64

Head follow 58

Eye and head controls

A single interaction layer keeps the first demo legible and responsive.

Eye follow

Tune how strongly the agent keeps your gaze in view.

Head follow

Adjust how much the neck and head join the eye response.

Use the arrows to move between intro and controls

Choose or import next

character choice

Pick a character or bring your own.

The second step repeats the app character switcher pattern: browse bundled characters, add a new profile, or drag a model into the import flow.

preset characters only

local session state

no cloud uploads or saved exports

Included characters

Start from a bundled profile, then promote the session when account persistence is available.

Import preview

Drag a GLB here or choose a model.

Trial mode can preview the import path, inspect metadata, and stage attribution before permanent upload is enabled.

Trial limitation: preset characters only, local session state, no cloud uploads or saved exports.

01 / Camera Follow Demo

Show one capability, not the whole workbench.

This section should become the first live demo panel. The agent is neck-up, the camera prompt is explicit, and the interface exposes only the two controls needed to explain gaze and posture.

What this section should communicate

Explain why this panel exists before asking for camera access.
Keep the controls local to this feature: eye follow intensity and head follow intensity.
Use the text block to explain what the audience should notice in the face and neck motion.

Section controls

Placeholder sliders for the focused feature UI.

camera behavior

Eye follow intensity

This should eventually drive iris motion and eye target locking.

Head follow intensity

Use this to explain how much neck and jaw should join the gaze response.

Jonathan bust studycamera follow

02 / Horizontal Story Rail

Use horizontal chapters only where they add structure.

Desktop can pin this section and move sideways through three or four cards. Tablet and mobile can still use the same cards, but the experience should remain comfortable and obviously scrollable.

What this section should communicate

Each card should feel like its own mini-demo with one agent and one point to notice.
Horizontal chapters are best for capability comparisons or short before/after narratives.
Keep the explanation short so the agent remains the main visual focus.

Slide 1

Eye + Head Follow

Horizontal panel one should isolate camera consent, subtle gaze follow, and a single explanation card about why neck-up presence feels social instead of robotic.

Demo focus

A single slider placeholder to reinforce the one-feature-at-a-time pattern.

Jonathan bust studystory slide 1

Slide 2

AI Face Variation

Panel two should show one character identity being tuned with a small UI: one slider for variation, one short paragraph about what the knob changes, and one visual state update.

Demo focus

A single slider placeholder to reinforce the one-feature-at-a-time pattern.

Jonathan bust studystory slide 2

Slide 3

Social Turn Taking

Panel three should stage one clean interaction cue at a time: waiting, listening, reacting, or handing the conversation off to a second agent.

Demo focus

A single slider placeholder to reinforce the one-feature-at-a-time pattern.

Jonathan bust studystory slide 3

03 / Face Variation

Reserve one section for identity and expression tuning.

This chapter should explain how AI-generated faces or preset identities are explored without overwhelming the page. It needs one clear visual change and one paragraph about what the audience is controlling.

What this section should communicate

Introduce face variation as a deliberate creative choice, not as a dense parameter panel.
Keep the explanation tied to practical outcomes such as warmth, age, polish, or stylization.
Pair the slider with a small caption that calls out exactly what changes in the face.

Face generation lab

Basic placeholder for the future AI face control UI.

identity shaping

Face variation

Higher values should push identity farther from the default portrait while staying on-brand.

Placeholder note: this is where we can explain whether the final implementation uses generated portraits, parameterized face styling, or swaps between curated identities.

Jonathan bust studyface variation

meet loom large

A calm first hello.

Tune the social response.

Pick a character or bring your own.

Drag a GLB here or choose a model.

Show one capability, not the whole workbench.

Use horizontal chapters only where they add structure.

Eye + Head Follow

AI Face Variation

Social Turn Taking

Reserve one section for identity and expression tuning.

End on interaction between agents, not a feature checklist.