Qwen Image Layered API Documentation

Qwen Image Layered API Documentation

Playground

Qwen-Image Layered is a unified image-layer decomposition model for prompt-guided compositing. Provide points, boxes, or rough masks to isolate subjects and regions, and the model splits a single image into multiple RGBA layers with clean alpha, soft edges, and correct occlusion order. Ready-to-use REST inference API with fast response, no cold starts, and affordable pricing.

Features

qwen-image/layered is a layered image decomposition model that splits a single image into multiple clean RGBA layers, enabling flexible compositing, background separation, and layer-based creative editing. It supports an optional prompt for better semantic separation and explicit control over the number of output layers.

🌟 Why it stands out

• Controllable layer count: choose exactly how many RGBA layers you want via num_layers.
• Clean RGBA outputs: each layer includes transparency for easy compositing and editing.
• Prompt-guided separation: optionally describe the scene to improve layer grouping in complex images.
• Workflow-friendly: ideal for design iteration, asset cleanup, and creative pipelines.

⚙️ Capabilities

• Image-to-multi-layer RGBA decomposition
• Transparent background handling per layer (RGBA output)
• Optional prompt conditioning for semantic grouping
• Works best on images with clear subjects, strong contrast, and limited heavy occlusion

⚙️ Parameters

💰 Pricing

Reference table:

How to use

1. Upload the source image (or provide a public URL).
2. Set num_layers to the number of layers you want.
3. Optional: add a prompt to improve semantic separation.
4. Run the model.
5. Download the RGBA layers and composite/edit them in your workflow.

💡 Best Use Cases

• Layer-based editing and compositing workflows
• Background separation and subject isolation
• Poster / banner / creative layout design
• Rapid asset preparation for marketing and social creatives

📝 Notes

• Best results: clear subjects, good lighting, minimal motion blur, and strong foreground/background separation.
• Higher num_layers can help break complex scenes into smaller components, but may also create finer splits that require selection/merging in post.

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",
  "num_layers": 4
}
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/wavespeed-ai/qwen-image/layered" \
  -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

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