Z Image Base LoRA

Z Image Base LoRA

Playground

Z-Image-Base LoRA (6B) enables high-quality text-to-image generation with full CFG support and external LoRA support. Supports applying up to 3 LoRAs for custom styles. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Z-Image Base LoRA is a 6-billion parameter text-to-image model from Tongyi-MAI with full LoRA support. Apply up to 3 custom LoRA adapters simultaneously to generate images with personalized styles, characters, or brand aesthetics — all while maintaining fast generation speeds.

Why Choose This?

• Triple LoRA support Apply up to 3 custom LoRA adapters at once for layered style control — combine character, style, and aesthetic LoRAs in a single generation.

• Flexible output sizing Customize width and height up to 1024px for any aspect ratio you need.

• Prompt Enhancer Built-in tool to automatically improve your prompts for better results.

• LoRA ecosystem compatibility Load LoRA weights from popular sources like Civitai and Hugging Face, or train your own custom LoRAs.

• Affordable pricing Just $0.012 per image — perfect for high-volume generation with custom styles.

Parameters

How to Use

1. Write your prompt — describe the image you want to create, including your LoRA trigger words.
2. Add LoRAs — click ”+ Add Item” to add up to 3 LoRA adapters with their weights.
3. Set dimensions — adjust width and height for your needs.
4. Run — submit and download your image.

Pricing

Best Use Cases

• Character Consistency — Use character LoRAs to maintain identity across multiple generations.
• Brand Aesthetics — Apply brand-specific style LoRAs for consistent marketing visuals.
• Art Style Transfer — Generate images in specific artistic styles trained into LoRAs.
• Combined Styles — Layer multiple LoRAs for unique style combinations.
• Rapid Iteration — Test different LoRA combinations quickly at low cost.

Pro Tips

• Include your LoRA trigger words in the prompt for best activation.
• Start with LoRA weight around 0.7-1.0, then adjust based on results.
• Combine complementary LoRAs (e.g., character + style + lighting) for richer outputs.
• Use the Prompt Enhancer to automatically improve your descriptions.
• Keep the same seed when comparing different LoRA combinations.

Train Your Own LoRA

Want to create custom LoRAs for Z-Image? Use the Z-Image LoRA Trainer:

• Z-Image LoRA Trainer — Train custom LoRAs with your own images.

Guidance

• How to Train your LoRA?
• How to Use your LoRA?

Related Models

• Z-Image Base — Base model without LoRA support at $0.01 per image.
• Z-Image Turbo — Faster generation optimized for sub-second inference.

Notes

• Maximum of 3 LoRAs can be applied per generation.
• LoRA weights typically range from 0.5 to 1.0 for best results.
• enable_sync_mode is only available through the API, not in the web interface.

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",
  "strength": 0.6,
  "seed": -1,
  "output_format": "jpeg"
}
JSON
)

# 1. Submit the prediction.
SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  -X POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/z-image/base-lora" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}")

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 "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id
' >&2
  exit 1
fi
RESULT_URL=$(printf '%s' "${TASK}" | jq -r '.urls.get // empty')
if [ -z "${RESULT_URL}" ]; then RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"; fi

# 2. Poll until the prediction finishes.
while true; do
  RESPONSE=$(curl --silent --show-error --fail-with-body \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}")
  RESULT=$(printf '%s' "${RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout) printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    created|processing) sleep 2 ;;
    *) printf 'Unexpected status: %s
' "${STATUS}" >&2; exit 1 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Response Parameters

Result Request Parameters

Result Response Parameters