Hidream O1 Image Dev Text To Image API Documentation

Hidream O1 Image Dev Text To Image API Documentation

Playground

HiDream-O1-Image Dev is a distilled, faster variant of the unified HiDream-O1-Image model — same native text-to-image / edit / personalization architecture at up to 2K, optimized for low-step inference and lower cost. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

HiDream-O1-Image Dev is the distilled variant of HiDream-O1-Image. It keeps the unified, single-native-model architecture (text-to-image, image editing, and subject-driven personalization) and the same up to 2K (≈2048×2048) resolution ceiling, but is tuned for fast, low-step inference — making it the best pick when you need speed and lower cost without giving up the HiDream-O1 look.

Why Choose This?

• Distilled for speed — Fewer denoising steps than the full model, which means lower latency per image.
• Lower cost — Half the price of the full HiDream-O1-Image model, ideal for high-volume generation, drafts, and rapid iteration.
• Same unified architecture — One native model handles text-to-image, edit, and personalization. No external adapters or pipelines.
• High-resolution output — Generate detailed images up to ~2048×2048 with strong composition and lighting.
• Strong prompt adherence — Faithfully follows complex, multi-clause descriptions including style, mood, and on-image text.
• Flexible aspect ratios — Supply any custom width/height up to the supported maximum (the model snaps to the closest supported resolution).
• Multiple output formats — JPEG, PNG, or WEBP.
• Reproducibility — Use a fixed seed to recreate exact results.

Parameters

How to Use

1. Write your prompt — describe subject, style, lighting, mood, and any on-image text.
2. Choose size — pick an aspect ratio that fits your use case (e.g., 2048*2048 for square, 1536*2048 for portrait, 2048*1152 for landscape).
3. Set output_format — JPEG for compact files, PNG for lossless, WEBP for balance.
4. Set seed (optional) — pin a number to reproduce exact results.
5. Run — submit the request and download the generated image.

Pricing

Flat rate per generated image — half the price of the full HiDream-O1-Image model.

Best Use Cases

• High-volume generation — Bulk content creation where per-image cost and latency matter.
• Drafts & rapid iteration — Quickly explore prompt variations before committing to the full model.
• Concept Art & Illustration — Translate text descriptions into detailed visual concepts.
• Marketing & Social Media — Produce eye-catching visuals for posts, ads, and campaigns.
• Storyboarding — Quickly visualize scenes and characters for creative projects.
• Product Visualization — Mock up product concepts and lifestyle imagery.

Dev vs. Full — Which to Pick?

Use Dev when you’re iterating, batching, or cost-sensitive. Switch to the full model for the very best fidelity on hero shots.

Pro Tips

• Be specific about subject, style, lighting, and composition — HiDream-O1 follows detailed prompts well.
• For posters or banners, describe layout and any on-image text explicitly.
• Pin a seed while iterating so you can A/B test prompt tweaks against a fixed baseline.
• Match size to your final use case (square for thumbnails, 16:9 for banners, 9:16 for mobile).
• Use png when you need lossless quality; use jpeg or webp to keep file sizes small.

Notes

• Only prompt is required — every other field is optional.
• Higher resolutions may slightly increase processing time.
• Ensure prompts comply with content guidelines.

Related Models

• HiDream-O1-Image Dev · Edit — Same distilled model, in image-editing mode.
• HiDream-O1-Image · Text-to-Image — The full (non-distilled) model for highest fidelity.
• HiDream-O1-Image · Edit — Full model in edit mode.

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": "2048*2048",
  "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/hidream-o1-image-dev/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