GPT Image 2 API

Introduction

GPT Image 2 Edit runs OpenAI's gpt-image-2 model on images you already have. Give it one to ten source images and an instruction: change a garment or a colour, retype a headline, relight a scene, or combine people and places from several sources into one frame. An optional PNG mask confines the repaint to a single area.

Choose a mode by starting material

Key features

Instruction editing, masked when needed

Without a mask the model works out from the prompt what to change and redraws the rest to match. With a mask, only the transparent pixels are repainted and everything else is kept.

Up to ten sources in one request

Refer to each source by its order: the first image supplies the person, the second the setting, the third the product. The model combines them into a single new frame.

Fidelity and shape control

Input fidelity high holds faces, fabric and printed details closer to the sources. Aspect ratio auto keeps the shape of the first source; any of fifteen fixed ratios gives the result that shape instead.

Request parameters

Pricing

Prices are in credits, the same credits you use on the website (1 credit = $0.01). The hold is taken when the task is accepted and settled to the actual charge when it finishes; failed tasks are refunded.

Each task is rounded up to whole credits, so one run at the lowest tier costs 1 credit per image ($0.01).

For an exact number before you run, send the same body to POST /jobs/quote.

Use cases

Retype and restyle finished artwork

Swap the words on a poster or cover, change a product colour or a garment, and keep the photograph underneath exactly as it was approved.

Composite scenes from references

Put a person from one photograph into the setting of another, or pair two people from separate shots, without rebuilding the layout from a text prompt.

Localised versions of one visual

Keep the photograph and replace the headline with a translated line for each market, one edit per language.

Prompt guide & pro tips

Say what stays as well as what changes

List the change first, then name what must not move: face, pose, lighting, background, logo. Anything you do not protect may be redrawn.

Number your sources

Write "the woman from the first image" and "the boat from the second image". The order of image_urls is the only way the model can tell them apart.

Use a mask for small fixes

For a single object or a patch of background, a PNG mask the same size as the first source is more reliable than describing the area in words.

Keep auto for faithful framing

Leave aspect ratio on auto when the edit should keep the framing of the source; pick a fixed ratio only when the result needs a new shape.

API quick start

Call the GPT Image 2 API in three steps

1. Create an API key. Sign in to PopNow and create a key in the API console. Keys start with sk-popnow- and use the same credits as the website.
2. Send the request. POST /jobs/createTask with model openai/gpt-image-2/edit and an input object that follows the parameter table. The response returns a task ID straight away.
3. Get the result. Poll GET /jobs/recordInfo with the task ID, or pass a callBackUrl and we post the result to it. Download the output links when the task succeeds.

# 1. Create the task. Keep the Idempotency-Key if you retry after a network error.
export IDEMPOTENCY_KEY="$(uuidgen)"
curl -X POST https://api.popnow.ai/api/v1/jobs/createTask \
  -H "Authorization: Bearer $POPNOW_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $IDEMPOTENCY_KEY" \
  -d '{
  "model": "openai/gpt-image-2/edit",
  "input": {
    "prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
    "image_urls": [
      "https://media.popnow.ai/influencer/examples/marco.webp"
    ],
    "aspect_ratio": "auto",
    "resolution": "1k",
    "quality": "medium",
    "num_outputs": 1,
    "input_fidelity": "low"
  }
}'

# 2. Copy data.taskId from the response, then poll until state is succeeded, failed, canceled or expired.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.popnow.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
  -H "Authorization: Bearer $POPNOW_API_KEY"

Replace the prompt and any links with your own. The API tab has the response, polling and callback reference.

GPT Image 2 API questions

All GPT Image models

GPT Image 2.5 Flare and Sunburst add transparent backgrounds, higher quality tiers and 16-photo edits.

• GPT Image 2.5 Sunburst APIOpenAI's quality-first model, up to 4K with clean textFrom 1 credit per image
• GPT Image 2.5 Flare APIOpenAI's speed-first model with masked edits and cut-outsFrom 1 credit per image
• GPT Image 2 APILegible in-image text and precise masked editsFrom 1 credit per image

Other image APIs

• Nano Banana 2.1 APIGoogle's newest Nano Banana, sharp text and 4K outputFrom 8 credits per image
• Nano Banana Pro APIPlans the layout first: dense text, diagrams and 4K editsFrom 17 credits per image
• Seedream 5.0 Pro APIHigh-detail images and edits from up to ten referencesFrom 6 credits per image
• FLUX.2 Max APITop-tier FLUX: photoreal light and exact brand colorsFrom 11 credits per image
• Qwen Image 2.1 Pro APIUp to six detailed 2K images from one promptFrom 6 credits per image
• Qwen Image 2.1 Turbo APIFast 8-step renders, up to six per promptFrom 3 credits per image

More about GPT Image 2

Try it

GPT Image 2 AI Image Generator onlineCreate images with GPT Image 2 in PopNow

Docs

GPT Image 2 model ID and input schema in the API docsPopNow API quickstart: your first request in five minutesImage generation guide for the PopNow APIWebhooks: get GPT Image 2 results without polling