Jib Mix Qwen Image Text To Image API Documentation

Jib Mix Qwen Image Text To Image API Documentation

Playground

Jib Mix Qwen is a next-gen Text-to-Image model optimized for producing natural, pretty faces with improved Asian facial rendering. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Jib-Mix-Qwen-Image is a finely tuned text-to-image generation model based on Qwen-Image 20B (MMDiT), optimized through the Jib-Mix portrait enhancement pipeline. It specializes in realistic human faces, cinematic lighting, and vivid artistic styles, delivering professional-grade visuals from simple text prompts — no LoRA setup needed.

Why it looks great

• Jib-Mix fine-tuning – Enhances facial structure, skin texture, and lighting realism, especially for close-ups and half-body portraits.
• Cinematic diffusion engine – Captures lifelike depth, atmosphere, and tone with consistent color harmony.
• Exceptional text rendering – Handles both Chinese and English typography natively, blending text naturally into the image.
• Broad style coverage – From photorealism to anime, oil painting, 3D, or stylized artwork—one model, infinite versatility.
• Identity consistency – Generates characters with coherent facial details and stable expressions across prompts.

Limits and Performance

• Max resolution per job: up to 1536 × 1536 pixels
• Output formats: JPEG / PNG / WEBP
• Processing speed: ~5–8 seconds per image (depending on prompt complexity)
• Prompt input: supports detailed, multi-line bilingual descriptions

Pricing

• $0.02 per image Each image is billed individually.

How to Use

1. Enter a prompt describing your desired image (Chinese or English).
2. Set image size (width × height, up to 1536×1536).
3. (Optional) Set a seed for reproducibility (-1 = random).
4. Choose output format (JPEG / PNG / WEBP).
5. Generate → preview → iterate with refined prompts.

Pro tips for best quality

• Be specific — describe lighting, pose, emotion, and background for more control.
• For portraits, include keywords like cinematic lighting, soft focus, 8K detail, professional photo.
• Fix seed to maintain subject consistency across multiple outputs.
• Experiment with styles (e.g., realistic, anime, oil painting, CG render) to explore model versatility.

Note

• For best realism, ensure prompts describe camera angle, lighting, and environment — the model responds strongly to cinematic cues.

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'
{
  "prompt": "A cinematic ocean wave at sunrise, highly detailed",
  "size": "1024*1024",
  "output_format": "jpeg"
}
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/jib-mix-qwen-image/text-to-image" \
  -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