Minimax H3 Singularity Image To Video LoRA API Documentation

Minimax H3 Singularity Image To Video LoRA API Documentation

Playground

MiniMax H3 Singularity Open Weights Image to Video animates a first-frame image, optionally with last-frame guidance, into coherent 480P / 540P / 768P videos with native stereo audio, 5-15 second duration, flexible aspect ratios, and per-second billing. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

MiniMax H3 Singularity Image-to-Video LoRA runs the open-weights edition of MiniMax H3 Singularity with LoRA support on WaveSpeedAI infrastructure. It combines first-frame image animation, optional last-frame guidance, native audio generation, and up to three LoRA weights in a single video generation workflow.

Why Choose This?

• Image-to-video generation
  Animate a first-frame image into a complete prompt-guided video.

• LoRA customization
  Load up to 3 LoRA weights per request, each with its own scale.

• Optional last-frame guidance
  Use last_image to guide the ending frame and create a more controlled start-to-end transition.

• Native audio generation
  Generate video and audio together in the same workflow.

• Prompt-guided sound direction
  Describe dialogue, music, ambience, and sound effects directly in the prompt.

• Four resolution tiers
  Generate at 480p, 540p, 768p, or 1080p.

• Flexible duration
  Generate clips from 3 to 15 seconds.

Parameters

How to Use

1. Upload the first-frame image — Provide the image that defines the opening frame, subject, composition, and visual direction.
2. Add last-frame guidance optional — Use last_image when the ending composition or pose matters.
3. Write the prompt — Describe the desired motion, camera movement, scene progression, timing, and audio direction.
4. Choose resolution — Select 480p, 540p, 768p, or 1080p.
5. Set duration — Choose a video length from 3 to 15 seconds.
6. Add LoRAs optional — Provide up to 3 LoRA files with individual scale values.
7. Set seed optional — Use a fixed seed when reproducibility is needed.
8. Submit — Generate the final video with audio.

Pricing

Pricing is based on generated video duration and selected resolution.

Best Use Cases

• LoRA-guided image animation — Animate a first-frame image while applying custom style, character, or concept LoRAs.
• Character animation — Bring portraits, characters, or illustrated subjects to life with prompt-guided motion.
• Start-and-end-frame control — Use image and last_image to guide both the beginning and ending of the video.
• Style customization — Apply LoRA weights to push the generated result toward a specific visual direction.
• Audio-driven scenes — Generate dialogue, ambience, music, or sound effects together with the video.
• Creative prototyping — Test different prompts, LoRA combinations, motion directions, and resolutions.

Pro Tips

• Use a clear first-frame image with strong subject visibility and composition.
• Use last_image when the final pose, framing, or scene state matters.
• Describe visible motion clearly instead of relying only on visual-style terms.
• Include camera movement and scene progression when motion direction matters.
• Keep LoRA combinations focused so different weights do not compete too strongly.
• Use 480p for lower-cost testing before moving to higher resolutions.
• Start with shorter durations while testing motion and LoRA behavior.
• Use a fixed seed when comparing LoRA scales or prompt changes.

Notes

• prompt and image are required.
• last_image is optional.
• duration supports 3–15 seconds.
• resolution supports 480p, 540p, 768p, and 1080p.
• Up to 3 LoRA weights can be used per request.
• LoRA loading does not add a separate charge.

Related Models

• MiniMax H3 Singularity Image-to-Video Open Weights — Animate a first-frame image with prompt-guided motion and audio.
• MiniMax H3 Singularity Reference-to-Video Open Weights — Generate video with image, video, and audio reference guidance.

Reference

• How to Train Your Own LoRA Model Without Coding
• How to Use Your LoRA on WaveSpeedAI

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",
  "image": "https://interactive-examples.mdn.mozilla.net/media/cc0-images/painted-hand-298-332.jpg",
  "resolution": "480p",
  "duration": 5
}
JSON
)

# 1. Submit the prediction.
SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  -X POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/minimax-h3-singularity/image-to-video-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 "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id
' >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 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|deleted) printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Response Parameters

Result Request Parameters

Result Response Parameters