Ideogram AI Ideogram V4.5 Edit API Documentation

Ideogram AI Ideogram V4.5 Edit API Documentation

Playground

Ideogram V4.5 Image Edit transforms images with written instructions, reference images, masks, and precision controls, supporting targeted edits, visual refinements, composition changes, and professional creative workflows. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Ideogram V4.5 Edit modifies an existing image using natural-language instructions, with optional reference images and mask-based control. Use it to change objects, colors, materials, text, layouts, and other visual elements while keeping the source image as the foundation.

Choose between regular editing and high-precision editing depending on how strongly unchanged areas need to be preserved.

Why Choose This?

• Prompt-based image editing
  Modify an existing image with natural-language instructions.

• Targeted visual changes
  Change objects, colors, materials, text, styling, and other image elements.

• Reference-image guidance
  Add supporting images to guide appearance, products, subjects, or visual direction.

• Mask-based editing
  Restrict changes to specific regions while preserving the rest of the image.

• High-precision mode
  Use edit_precision=high when preserving unaffected image regions is especially important.

• Four quality tiers
  Choose very_low, low, medium, or high depending on cost and output requirements.

Parameters

How to Use

1. Upload the source image — Provide the image you want to modify.
2. Write the edit instruction — Describe what should change and what should remain unchanged.
3. Add reference images optional — Supply supporting images when appearance or style guidance is useful.
4. Add a mask optional — Restrict editing to specific regions when precise localization is needed.
5. Choose edit precision — Use regular for standard editing or high when preservation of unaffected details matters more.
6. Choose quality — Select very_low, low, medium, or high.
7. Submit — Generate and retrieve the edited image.

Pricing

Pricing is based only on the selected quality tier.

The default medium quality edit costs $0.06 per image.

edit_precision, size, mask_url, and the number of reference_images do not add separate charges.

Best Use Cases

• Product variations — Change product colors, materials, finishes, or visual styling.
• Advertising edits — Revise layouts, graphics, text, and campaign imagery.
• Selective corrections — Modify a specific area while preserving the rest of the source image.
• Interior styling — Change furniture, materials, colors, or decorative elements.
• Typography edits — Revise text or designed visual elements inside an existing image.
• Reference-guided editing — Use supporting images to guide products, subjects, or style changes.

Pro Tips

• State the requested change directly instead of redescribing the entire source image.
• Mention what should remain unchanged when preservation matters.
• Use edit_precision=high when unaffected regions need stronger preservation.
• Use a mask when the edit should be limited to a specific part of the image.
• Keep reference images closely related to the requested change.
• Use very_low or low quality for rapid iteration before moving to higher-quality output.
• Check text, small details, and masked boundaries before publishing the final result.

Notes

• Custom sizes are supported only for regular edits without a mask. Dimensions must be at least 256px and multiples of 32, with a maximum of 4,194,304 pixels and an aspect ratio no greater than 6:1.
• Leave size empty to follow the source image.
• For high-precision or masked edits, leave size unset.
• Masks must match the source image dimensions and contain both black areas to edit and white areas to preserve.
• Up to 4 reference images are supported, or 3 when using a mask. Only the source image is edited.
• Large input images may be downscaled.

Related Models

• Ideogram V4.5 — Generate images from text with typography, composition, and prompt control.
• Ideogram V4.5 Edit — Edit existing images with prompts, references, masks, and precision control.

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

set -euo pipefail

export WAVESPEED_API_KEY="your-api-key"

REQUEST_BODY=$(cat <<'JSON'
{
  "image": "https://interactive-examples.mdn.mozilla.net/media/cc0-images/painted-hand-298-332.jpg",
  "prompt": "A cinematic ocean wave at sunrise, highly detailed",
  "aspect_ratio": "1:1",
  "resolution": "1k",
  "edit_precision": "regular",
  "quality": "medium"
}
JSON
)

# 1. Submit the prediction once. Never retry this request automatically:
# a retry after an unclear failure can create a second, billed task.
CHECK='The task may already exist. Check History or the predictions list before submitting again.'
if ! SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --connect-timeout 10 --max-time 60 \
  -X POST "https://api.wavespeed.ai/api/v3/ideogram-ai/ideogram-v4.5/edit" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}"); then
  printf 'Submit failed. %s\n' "${CHECK}" >&2
  exit 1
fi

TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id. %s\n' "${CHECK}" >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction reaches a final status.
BODY_FILE=$(mktemp)
trap 'rm -f "${BODY_FILE}"' EXIT
while true; do
  # Transient errors (429, 5xx) are retried, honoring Retry-After.
  HTTP_CODE=$(curl --silent --show-error --fail-with-body \
    --connect-timeout 10 --max-time 60 --retry 5 --retry-delay 2 \
    --output "${BODY_FILE}" --write-out '%{http_code}' \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}" || true)
  case "${HTTP_CODE}" in
    2??) ;;
    000|429|5??)  # dropped connection or temporary error: keep polling
      printf 'Task %s: HTTP %s, polling again in 10s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      sleep 10; continue ;;
    *) printf 'Task %s: HTTP %s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      cat "${BODY_FILE}" >&2; exit 1 ;;
  esac
  RESULT=$(jq 'if type == "object" and has("data") then .data else . end' "${BODY_FILE}")
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout|deleted)
      printf 'Task %s ended with status %s\n' "${PREDICTION_ID}" "${STATUS}" >&2
      printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Conditional constraints

• Not allowed: reference_images has at least 4 items and mask_url is provided.

Every model also accepts the global request options enable_sync_mode and enable_base64_output; see their pages for details.

Response Parameters

Result Request Parameters

Result Response Parameters