muapi Logomuapi Logo
muapi Logomuapi Logo

HappyHorse 1.0 Reference to Video: 画像から動画生成

Happy Horse 1.0 Reference to Video generates video from 1-9 reference images, with selectable 720p/1080p resolution. Happy Horse 1.0 Reference to Video - generate expressive video clips conditioned on 1-9 reference images plus a text prompt. Selectable output resolution (720p/1080p) - price scales with the resolution chosen.

Interactive model controls

Happy Horse 1.0 Reference to Video - generate expressive video clips conditioned on 1-9 reference images plus a text prompt. Selectable output resolution (720p/1080p) - price scales with the resolution chosen.

📝

概要

このモデルについて

更新日: 2026年9月11日

Happy Horse 1.0 Reference to Video generates expressive video clips conditioned on 1-9 reference images plus a text prompt, at your choice of 720p or 1080p output resolution. Reference multiple images in your prompt (e.g. "@image1 riding @image2") to compose scenes that combine characters, objects, and settings from separate source images.

A single resolution parameter lets you draft cheaply at 720p and switch to 1080p for final output on the same endpoint.

1Multi-subject composition: combine a character and an object/vehicle from separate reference images into one scene
2Character-consistent video: keep a character's appearance consistent across a generated video using a reference photo
3Product-in-scene video: place a product reference image into a generated environment
4Concept mashups: blend visual elements from multiple references into a single coherent video
5Marketing creative: generate branded video content referencing product/character assets
💰

料金と価値

コスト分析

muapiapp$0.21–$0.42/秒 (720p–1080p)

Simple per-second, resolution-scaled 料金 with サブスクリプション不要s. Pick 720p for cost-effective drafts or 1080p for premium output, on the same endpoint.

** 競合サービスの料金は類似のモデル構成および利用ティアに基づいて算出された推定値です。

⚙️

技術詳細

設定スキーマ

プロンプトstring

Text description of the desired video. Up to 5,000 non-Chinese (or 2,500 Chinese) characters.

デフォルト値Use @image1 riding inside @image2 at extreme speed through a massive supermarket. The rocket cart blasts through aisles, launches over checkout counters, and sends products exploding everywhere while the camera chases closely behind through the chaos.
解像度Enum（2個の選択肢）

Output resolution. Price scales with resolution: 720p is cheaper, 1080p is more expensive.

デフォルト値720p
Reference 画像sarray

処理の開始点となる入力元画像のURLまたはBase64データ。

デフォルト値https://d3adwkbyhxyrtq.cloudfront.net/webassets/videomodels/happy-horse-1-reference-to-video-720p-1.jpg
アスペクト比Enum（5個の選択肢）

生成メディアのアスペクト比（例：16:9、9:16、1:1など）。

デフォルト値16:9
長さ（秒）int

生成する動画または音声の長さ（秒単位）。

デフォルト値5
シード値int

生成の再現性を保つためのランダムシード値。同じ値を使用すると同じ結果が得られます。

デフォルト値0
📖

実装ガイド

開発者ドキュメント

  1. Provide 1-9 reference image URLs in images_list (JPEG/PNG/WEBP, ≥400px shortest side, ≤10MB each).
  2. Write a prompt referencing the images by position (e.g. "@image1 riding @image2 through a city").
  3. Set resolution to "720p" (cheaper, default) or "1080p" (premium output).
  4. Select aspect ratio, duration (3-15s), and optional seed for reproducibility.
  5. Submit your request—the model composes a video from your references at the selected resolution.
❓

よくある質問

FAQ

What is the cost of Happy Horse 1.0 Reference to Video?

Happy Horse 1.0 Reference to Video costs $0.21/sec at 720p (default) or $0.42/sec at 1080p on muapi. A 5-second 720p clip costs $1.05; the same clip at 1080p costs $2.10.

How do I choose the resolution?

Pass `resolution: "720p"` (default) or `resolution: "1080p"` in your request body — both are served by the same endpoint.

How many reference images can I use?

1 to 9 images. Reference each one in your prompt via @image1 through @image9, matching the order in `images_list`.

Can I use a fixed seed for reproducibility?

Yes, pass an integer `seed` (0-2147483647) to get more consistent results across runs with the same prompt/references.

What happened to the dedicated 720p/1080p endpoints?

They still work exactly as before, unchanged, for existing integrations. This unified endpoint is the new recommended way to access both resolutions through one integration.