Prefect Pony Xl LoRA API Documentation

Prefect Pony Xl LoRA API Documentation

Playground

High-quality anime-style images from Prefect Pony XL with LoRA support for personalized characters and styles. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Prefect Pony XL LoRA is an anime-focused text-to-image model with on-the-fly LoRA loading. It keeps the vibrant, stylized look of Prefect Pony XL and lets you apply up to 3 LoRAs in a single request for consistent characters, signature art styles, or specific concepts.

Why Choose This?

• Anime and illustration quality: Tuned for character art, expressive faces, clean linework, and bold color.
• LoRA support: Load up to 3 LoRAs per request, each with its own scale, for targeted style and subject control.
• Custom dimensions: Set any width and height to match your target format.
• Prompt adherence control: Adjust guidance_scale to trade off between creativity and strict prompt following.
• Reproducible results: Fix the seed to recreate an image or explore controlled variations.
• Multiple output formats: Export as JPEG, PNG, or WebP.

Parameters

How to Use

1. Write your prompt — describe the subject, style, colors, and mood. Tag-style prompts work well.
2. Set dimensions — choose the width and height for your target format.
3. Add LoRAs (optional) — provide up to 3 LoRAs, each with its own path and scale.
4. Adjust guidance scale (optional) — raise it for stricter prompt following, lower it for looser, more creative results.
5. Set seed (optional) — use -1 for random, or a fixed number to reproduce a result.
6. Choose output format — JPEG for smaller files, PNG for lossless quality, WebP for a balance of both.
7. Run — submit the request and download your image.

Pricing

Flat rate per image generation. Adding LoRAs does not change the price.

Best Use Cases

• Consistent characters — apply a character LoRA to keep the same look across many images.
• Signature art styles — use a style LoRA to match a specific illustrator look or series aesthetic.
• Anime portraits and key art — generate expressive character portraits, covers, and posters.
• Concept and fan art — explore outfits, poses, and scenes quickly and affordably.
• Stickers and social content — produce eye-catching stylized visuals at volume.

Pro Tips for Best Results

• Use LoRAs trained for Pony-family SDXL models. LoRAs built for other base models will not load correctly.
• Include the LoRA’s trigger words in your prompt when the LoRA requires them.
• Start with a LoRA scale around 0.7–1.0. Lower it if the style overpowers your prompt, and raise it if the effect is too subtle.
• When stacking multiple LoRAs, reduce each scale slightly so they do not compete.
• Quality tags such as score_9, score_8_up, score_7_up at the start of the prompt often improve results.
• Keep the seed fixed while tuning LoRA scale or guidance so you can compare changes directly.

Notes

• The LoRA path must be a publicly accessible direct download URL to a .safetensors file.
• Generation time may vary with resolution, the number of LoRAs, and current queue load.
• For generation without LoRAs, see Prefect Pony XL.

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",
  "guidance_scale": 3.5,
  "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/prefect-pony-xl-lora" \
  -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