Ultra Real Photo · skill Open Skill công khai

Ultra Real Photo

Q: Ask: My after-hours shots need one less digital-polished look. A: Get: I use Ultra Real Photo to turn a written scene into a photorealistic image. Q: Ask: Can I control the frame for a creepy stairwell shot? A: Get: Yes. Portrait is the default at 3:4. I can request landscape or square instead. Q: Ask: What goes in the description? A: Get: I add the people, setting, lighting, and other details I want included.

@MayaCalloway

Custom

2 1 2

https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/550336/2026-10-15/46wxmc9u.webp

# Ultra Real Photo

Describe it, and it's rendered as a photograph you'd believe was real.

## When to use

- The user wants a photorealistic image generated from a text description (hyper-realistic photo, lifelike scene)

Not for: Editing an existing photo (use an image-edit skill).

## Inputs and gates

| Input | Required | Default | Gate |
|---|---|---|---|
| Description | Yes | — | What the user wants; merged into the prompt |
| Orientation | No | `portrait default` | portrait default; landscape or square on request |

- Fashion, dark and horror themes are fine as stills; refuse explicit sexual content and anything sexualizing minors.

## Steps

1. Expand the user's idea into the brief below. If they gave almost nothing, use the tuned defaults instead of asking.
2. `generate_image(prompt=<PROMPT>, model='t2i.hd', orientation='portrait', aspect_ratio='3:4', name='Ultra Real Photo', entity_type='skill_generation')`.
3. **Self-check, one retry**: if the composition is cluttered or off-brief, regenerate once with a simpler subject and clearer light.
4. Reply in one or two sentences, e.g. "All set. Your ultra real photo is on the canvas. Next I can do landscape." Results appear on the canvas; never paste URLs.

## Prompt templates

**PROMPT**

```text
Subject: <who/what>. Action: <what they do>. Location: <where>. Composition: <shot size, angle, lens>. Style: an ultra photorealistic photograph of <description>, natural lighting, real textures, camera-like depth of field. Light: <key light, mood>. No text, no logos.
```

## Failure handling

- `success=false`: say what failed in one sentence; retry once only if transient; never invent links.

## Examples

**Default**: the user says "this one". Use orientation `portrait default`, 3:4. Reply: "All set. Your ultra real photo is on the canvas."
**Variation**: "next variation". Keep the same inputs and change one option (landscape).
**Thin request**: the user gives one word. Fill the brief from the tuned defaults and generate; do not interrogate.