Files
hypit/packages/gpt-image-kits/README.md
kashorin c54b43f94a refactor(image): align prompt kits and craft around visual intent
Separate Capture, Person, Shot and Setting in the phone-image Kit and update all 12 example callers. Clarify reference interpretation, Gen-Z styling language, explicit portrait anchors and format-specific camera relationships. Replace the portrait studies with the accepted four-paragraph directions and connect the relevant Skill routes.

Validation: three example Sources pass hypit check; all 12 prompts and a reference-derived optional-slot case assemble through the Text renderer; Skill validation and local documentation links pass.
2026-09-10 10:56:09 +08:00

2.4 KiB

@hypit/gpt-image-kits

phone-ugc-v1 is a data-only Text Template for assembling image prompts. Its source emits a fixed iPhone-video capture paragraph, followed by the supplied Text slots in person → shot → setting order, with paragraph separators. Import it as @hypit/gpt-image-kits/phone-ugc-v1.

For creative direction, see the Image direction page in the Hypit Skill.

Blocks and inputs

The capture block is fixed by this template version. Selecting a different capture language means selecting or authoring another Text Template. The caller supplies the variable direction:

Text slot Required Content
person No Person or cast, appearance and styling; identity references and what they preserve.
shot Yes Camera view, framing, posture, gaze, action and interaction with people or props.
setting No Surrounding place, palette, structures and objects; scene references and what they preserve.

For a complete new character-and-scene image, supply all three as ordinary paragraphs. A derived view may inherit its person or setting from connected references and omit those slots. Omitted blocks add no empty paragraphs. State the necessary reference responsibilities in the slot they affect. Each slot accepts an ordinary Text Output and can describe more than one person or relationship. The template produces Text; the image Surface consumes it. Actual reference Resources and model parameters are connected on that Surface.

Assemble the prompt

Given authored Text Outputs portrait-person, portrait-shot and portrait-setting, this fragment assembles the prompt and passes it to gpt:Image:

<import as="text" from="@hypit/text@1"/>
<import as="gpt" from="@hypit/gpt-image@1"/>
<import as="ugc" source="@hypit/gpt-image-kits/phone-ugc-v1"/>

<text:Render id="portrait-prompt" template={ugc.phone-ugc-v1}>
  <text:Set name="person" text={portrait-person}/>
  <text:Set name="shot" text={portrait-shot}/>
  <text:Set name="setting" text={portrait-setting}/>
</text:Render>

<gpt:Image id="portrait" prompt={portrait-prompt} aspect-ratio="9:16" resolution="2K"/>

Connect reference images through gpt:Reference children on gpt:Image in the order described by the prompt. For example, a person paragraph can inherit identity from reference 1 while shot describes holding the product from reference 2. The Text does not create Resource bindings.