Flux Schnell LoRA API Documentation

Flux Schnell LoRA API Documentation

Playground

FLUX.1 [schnell] is a 12B-parameter rectified flow transformer that generates images from text prompts and supports LoRA adapters. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

wavespeed-ai/flux-schnell-lora is a fast, high-throughput text-to-image model with LoRA support for style and identity control. Use it when you want Schnell-level speed, but also need consistent aesthetics across batches (brand looks, character styles, “same vibe” series).

Key capabilities

• LoRA-guided generation for consistent style, mood, or subject traits
• Ultra-fast iterations for prompt exploration and batch workflows
• Optional image + mask inputs for reference-based edits (when enabled)
• Up to 3 LoRAs per run (mix styles or add a single strong look)

Inputs

• Prompt (required): text description of the image you want

• Image (optional): reference image input (if available in your deployment)

• Mask image (optional): mask for local edits / inpainting (if available)

• Strength (optional): how strongly edits follow the image/mask intent

• LoRAs (optional): up to 3 LoRA items, each with:

• Path: LoRA identifier (for example, a LoRA repo/name)

• Scale: LoRA weight/strength

• Size: output width and height

• Num images: number of images to generate

• Seed: random seed for reproducibility

• Output format: jpeg, png, etc.

• Base64 output (API only): return base64 instead of a URL

• Sync mode (API only): wait for the result before returning

Recommended usage patterns

• One-LoRA “signature look”: start with a single LoRA and moderate scale, then iterate prompts for composition and lighting.
• Style blending: combine 2 LoRAs (e.g., lighting + texture) and keep scales conservative to avoid overcooking details.
• Consistency runs: fix the seed, lock the LoRA set, then vary only one prompt element per iteration.

Use cases

• Brand creative at scale: generate a whole campaign’s visuals in one consistent style
• Character/style series: keep the same “feel” across thumbnails, posters, or profile sets
• Product mock concepts: explore layouts quickly while preserving a chosen aesthetic
• Rapid A/B exploration: batch variations under a stable LoRA look for easy comparison
• Creator workflows: build a repeatable visual identity without sacrificing speed

Pricing

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",
  "strength": 0.8,
  "size": "1024*1024",
  "num_images": 1,
  "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/flux-schnell-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