FREE TO EXPLORE · BUILT FOR PRACTICAL RESULTS

A clearer prompt.
A more useful result.

Find your task. Make the prompt your own. Copy it into ChatGPT, then check the result with a clear standard.

1200 prompts availableEnglish · Español · Português (BR)

Explore by task

Start here

  1. Choose a task and read its upload instructions.
  2. Edit the fields, then copy the complete prompt.
  3. Paste into ChatGPT. Upload source material there when requested.
  4. Compare the result with the example and checks. Refine one change at a time.

These are natural-language prompts, not official slash commands. A label such as /cinematicportrait does not activate a hidden ChatGPT feature. Tool access and usage limits depend on your account. AstraCruz does not execute prompts in your account. Image edits can change faces and product details; restorations may invent missing content. Check downloaded pixel dimensions: a prompt does not guarantee 4K or print-ready resolution.

Customization happens in this browser. No account or payment is required here. Do not put passwords or confidential material in prompts. These templates are editorially reviewed; they have not all been individually executed in ChatGPT.

Your prompt library

Portraits & natural retouchingCinematic portrait

Purpose

Shape the light into a soft side key, lower background saturation, and add restrained film contrast. Retain natural skin tone and facial structure.

How to use it

Upload the original photograph you have permission to use in ChatGPT. Keep an untouched copy. Identify the area to change; provide a second labeled image when the task requires a background or reference.

Your ready-to-copy prompt

Work with the uploaded image. Shape the light into a soft side key, lower background saturation, and add restrained film contrast. Retain natural skin tone and facial structure.

Additional details: {{details}}
Intended use: {{format}}

Change only what this task requires. Preserve all other visible content as closely as possible. If a required reference is missing, ask before editing. Do not claim exact identity preservation or recovered missing information.

Acceptance check: Eyes remain visible; skin is not orange; shadows retain detail.

Make it your own

  • Subject, area, palette, or reference details: Use the uploaded source and the task description; ask if a required detail is missing.
  • Intended use and framing: Personal review; preserve the original framing unless this task requires another format.

What a successful result should contain

A successful result should meet this check: Eyes remain visible; skin is not orange; shadows retain detail.

Improve the result

Compare the original and result side by side. If this check fails, name the exact area and request one correction at a time: Eyes remain visible; skin is not orange; shadows retain detail.

Limits & checks

Generative edits can change facial identity, anatomy, age, or fine detail. Compare against the original at full size; reject identity drift.

Portraits & natural retouchingExecutive headshot

Purpose

Create a restrained corporate headshot with a neutral gray background, balanced face lighting, and a chest-up crop. Keep existing clothing and expression.

How to use it

Upload the original photograph you have permission to use in ChatGPT. Keep an untouched copy. Identify the area to change; provide a second labeled image when the task requires a background or reference.

Your ready-to-copy prompt

Work with the uploaded image. Create a restrained corporate headshot with a neutral gray background, balanced face lighting, and a chest-up crop. Keep existing clothing and expression.

Additional details: {{details}}
Intended use: {{format}}

Change only what this task requires. Preserve all other visible content as closely as possible. If a required reference is missing, ask before editing. Do not claim exact identity preservation or recovered missing information.

Acceptance check: Both shoulders fit; no invented tie or accessories; facial proportions match.

Make it your own

  • Subject, area, palette, or reference details: Use the uploaded source and the task description; ask if a required detail is missing.
  • Intended use and framing: Personal review; preserve the original framing unless this task requires another format.

What a successful result should contain

A successful result should meet this check: Both shoulders fit; no invented tie or accessories; facial proportions match.

Improve the result

Compare the original and result side by side. If this check fails, name the exact area and request one correction at a time: Both shoulders fit; no invented tie or accessories; facial proportions match.

Limits & checks

Generative edits can change facial identity, anatomy, age, or fine detail. Compare against the original at full size; reject identity drift.

Portraits & natural retouchingTemporary blemish retouch

Purpose

Reduce only the temporary blemishes I identify, preserving pores, freckles, permanent marks, and age-related texture.

How to use it

Upload the original photograph you have permission to use in ChatGPT. Keep an untouched copy. Identify the area to change; provide a second labeled image when the task requires a background or reference.

Your ready-to-copy prompt

Work with the uploaded image. Reduce only the temporary blemishes I identify, preserving pores, freckles, permanent marks, and age-related texture.

Additional details: {{details}}
Intended use: {{format}}

Change only what this task requires. Preserve all other visible content as closely as possible. If a required reference is missing, ask before editing. Do not claim exact identity preservation or recovered missing information.

Acceptance check: Skin retains visible texture; permanent marks remain.

Make it your own

  • Subject, area, palette, or reference details: Use the uploaded source and the task description; ask if a required detail is missing.
  • Intended use and framing: Personal review; preserve the original framing unless this task requires another format.

What a successful result should contain

A successful result should meet this check: Skin retains visible texture; permanent marks remain.

Improve the result

Compare the original and result side by side. If this check fails, name the exact area and request one correction at a time: Skin retains visible texture; permanent marks remain.

Limits & checks

Generative edits can change facial identity, anatomy, age, or fine detail. Compare against the original at full size; reject identity drift.

Portraits & natural retouchingReduce facial shine

Purpose

Soften oily highlights on the forehead, nose, and cheeks without flattening the face or darkening the rest of the photograph.

How to use it

Upload the original photograph you have permission to use in ChatGPT. Keep an untouched copy. Identify the area to change; provide a second labeled image when the task requires a background or reference.

Your ready-to-copy prompt

Work with the uploaded image. Soften oily highlights on the forehead, nose, and cheeks without flattening the face or darkening the rest of the photograph.

Additional details: {{details}}
Intended use: {{format}}

Change only what this task requires. Preserve all other visible content as closely as possible. If a required reference is missing, ask before editing. Do not claim exact identity preservation or recovered missing information.

Acceptance check: Highlights are controlled while cheek and nose contours remain.

Make it your own

  • Subject, area, palette, or reference details: Use the uploaded source and the task description; ask if a required detail is missing.
  • Intended use and framing: Personal review; preserve the original framing unless this task requires another format.

What a successful result should contain

A successful result should meet this check: Highlights are controlled while cheek and nose contours remain.

Improve the result

Compare the original and result side by side. If this check fails, name the exact area and request one correction at a time: Highlights are controlled while cheek and nose contours remain.

Limits & checks

Generative edits can change facial identity, anatomy, age, or fine detail. Compare against the original at full size; reject identity drift.

Portraits & natural retouchingGentle under-eye correction

Purpose

Lighten deep under-eye shadows modestly while retaining eyelid shape, fine lines, and natural differences between the two eyes.

How to use it

Upload the original photograph you have permission to use in ChatGPT. Keep an untouched copy. Identify the area to change; provide a second labeled image when the task requires a background or reference.

Your ready-to-copy prompt

Work with the uploaded image. Lighten deep under-eye shadows modestly while retaining eyelid shape, fine lines, and natural differences between the two eyes.

Additional details: {{details}}
Intended use: {{format}}

Change only what this task requires. Preserve all other visible content as closely as possible. If a required reference is missing, ask before editing. Do not claim exact identity preservation or recovered missing information.

Acceptance check: The subject still looks their age; neither eye changes shape.

Make it your own

  • Subject, area, palette, or reference details: Use the uploaded source and the task description; ask if a required detail is missing.
  • Intended use and framing: Personal review; preserve the original framing unless this task requires another format.

What a successful result should contain

A successful result should meet this check: The subject still looks their age; neither eye changes shape.

Improve the result

Compare the original and result side by side. If this check fails, name the exact area and request one correction at a time: The subject still looks their age; neither eye changes shape.

Limits & checks

Generative edits can change facial identity, anatomy, age, or fine detail. Compare against the original at full size; reject identity drift.

Portraits & natural retouchingEyeglass glare reduction

Purpose

Reduce reflections covering the eyes in the glasses. Preserve the frames and existing visible eye detail; do not invent hidden iris detail.

How to use it

Upload the original photograph you have permission to use in ChatGPT. Keep an untouched copy. Identify the area to change; provide a second labeled image when the task requires a background or reference.

Your ready-to-copy prompt

Work with the uploaded image. Reduce reflections covering the eyes in the glasses. Preserve the frames and existing visible eye detail; do not invent hidden iris detail.

Additional details: {{details}}
Intended use: {{format}}

Change only what this task requires. Preserve all other visible content as closely as possible. If a required reference is missing, ask before editing. Do not claim exact identity preservation or recovered missing information.

Acceptance check: Frame geometry is unchanged; unavailable eye detail is identified rather than presented as recovered.

Make it your own

  • Subject, area, palette, or reference details: Use the uploaded source and the task description; ask if a required detail is missing.
  • Intended use and framing: Personal review; preserve the original framing unless this task requires another format.

What a successful result should contain

A successful result should meet this check: Frame geometry is unchanged; unavailable eye detail is identified rather than presented as recovered.

Improve the result

Compare the original and result side by side. If this check fails, name the exact area and request one correction at a time: Frame geometry is unchanged; unavailable eye detail is identified rather than presented as recovered.

Limits & checks

Generative edits can change facial identity, anatomy, age, or fine detail. Compare against the original at full size; reject identity drift.

Natural-language prompt templates

See an actual edit

Demonstration created with the image-generation tool in ChatGPT Work on October 29, 2026. The source mug is fictional. The second image was produced by editing that source, not by generating an unrelated comparison. This is one example, not a promise of identical results.

AI-generated blue mug on a wooden desk in a kitchen
Before · AI-generated source
The same source mug edited onto a gray studio background
After · edited from that source
Replace the kitchen and wooden desk with a seamless light gray studio backdrop. Match the original left-side light and retain a plausible contact shadow. Keep the blue mug, handle, rim, proportions and camera angle as unchanged as possible.

Review focus: backdrop replacement, contact shadow, mug silhouette, handle and rim. Small changes in reflections or shape can still occur; exact product preservation is not guaranteed.

Transparency: a useful result with visible limits

This second edit used the same generated source. The downloaded PNG was checked for a real alpha channel, including transparent pixels. The tool changed the framing and some edge detail. The checkerboard below is only a display background, not part of the image. Inspect the handle opening and edge fringe before reuse.

AI-generated blue mug on a wooden desk in a kitchen
Before · AI-generated source
Actual transparent cutout of the generated blue mug; framing differs from source
After · edited from that source
Remove the kitchen and desk. Keep only the blue mug, with genuine transparency outside it and inside the handle. Preserve its shape and glaze as closely as possible.

ASTRACRUZ AI

Keep building your skills

Use the prompts, then go deeper with practical lessons, books, or a conversation with the community.

Capability references

OpenAI · ChatGPT ImagesOpenAI · Image inputs FAQOpenAI · ChatGPT capabilities