happyhorse-1.0-video-edit is the video editing model of HappyHorse 1.0, designed for style transfer, outfit changes, and element replacement in existing clips. Based on the original video and text editing intent, it can incorporate reference images to indicate the target appearance and optionally preserve the original audio. It is suitable for workflows that already have footage and want to create visual variations rather than generate entirely new videos from scratch.
Submit requests to the public API at api.acedata.cloud using the documented parameters, then use the results in your application.
api.acedata.cloudhappyhorse-1.0-video-editInput parameters and result formats vary by service. Use the public API for this model and follow its guide for generation, task retrieval and editing operations.
Resolution, duration, frame rate, and format are native specifications; reference images, audio options, and task delivery are used according to this platform's video editing workflow.
Use an existing video as the editing foundation and describe in text the style, clothing, or visual elements that need to change. Prompts should specify both what to modify and what to preserve, for example, changing the clothing material while retaining the original camera movement, so the task focuses on transforming the footage rather than redesigning the entire shot.
When text cannot accurately describe textures, color schemes, or clothing details, reference images can help convey them. Video editing supports up to five images, making them suitable for providing examples of target clothing, materials, or styles; reference images guide the direction of modification and should not be understood as pixel-by-pixel copying or strict local locking.
Visual transformation does not necessarily require replacing sound. For clips with existing narration, music, or ambient audio, use origin to preserve the original video audio; auto can also be selected. Before delivery, inspect visual modifications and audio results separately to avoid assuming that the sound will change according to the same intent simply because the visual style has changed.
Input an already filmed product clip, then provide target material or visual style images, and use text to describe the desired color palette and decorative direction. The output can be used to compare clip variations across different creative concepts, making it suitable for marketing previews with existing footage; product branding and key appearance details still need frame-by-frame review.
Use a person clip as input, pair it with target clothing images, and describe the apparel to replace and the shot content that needs to be preserved. The generated result is suitable for discussing styling directions or creating concept samples; pay close attention to turns, occlusions, and rapid movement, and do not treat the preview result directly as an exact apparel reproduction.
Submit visual editing instructions for clips with existing narration or location sound, and select origin to preserve the original audio. Applications can receive task IDs asynchronously, retrieve video links through queries or callbacks, and then proceed to review and editing workflows, making this suitable for material where audio content is already finalized and only the visual style needs adjustment.
This model starts with a video to be edited. If you only have a script, choose happyhorse-1.1-t2v; if you only have a first-frame image, consider happyhorse-1.1-i2v; if you want to generate new shots from reference images, consider happyhorse-1.1-r2v. The 1.1 generation models do not automatically replace 1.0 video editing; the choice should be based on the input material and operational goal.
When the goal is style transfer, outfit changes, or element replacement, this model directly matches the task. If the core requirement is to recreate the effects or camera movement of another video, consider wan2.7-videoedit. The two should not be judged as better or worse based solely on version numbers; preserving existing shots while changing appearance and transferring the motion effects of another video are different editing goals.
Prepare the source video video_url, editing instructions, and optional reference image_urls; select audio_setting=origin when audio needs to be preserved.
Set model=happyhorse-1.0-video-edit and action=video_edit for /happyhorse/videos. The editing duration follows the source video and does not arbitrarily trim or extend it using the shared duration.
Use async or callback_url to integrate background tasks, save the task_id, and query /happyhorse/tasks; wait for succeeded before reading video_url, then check people, actions, and audio.
Keep the people, actions, camera angle, and dialogue from the original video, changing only the jacket to the dark green style in the reference image, with matching shadows and lighting.
Use video_edit, video_url, and reference image_urls; retain the original audio setting audio_setting=origin, and check lip sync, clothing edges, and shot duration.
You cannot submit only text for video editing. This model requires the video to be edited video_url and the editing instruction prompt; reference images are optional supporting materials. If you want to generate a short video directly from text, you should use the t2v model; if you only have a static first-frame image, you should choose the i2v model.
This model's editing operation supports zero to five reference images, submitted through image_urls. Nine reference images fall within the scope of the reference-image-to-video operation and do not apply here. It is recommended to select images directly related to the intended modification and explain their purpose in the prompt to avoid mixing conflicting styles.
Set audio_setting to origin to preserve the original video audio; the other available value is auto. If the task only changes the visual appearance and the original narration or background music still needs to be retained, origin better fits this goal. After completion, you should still preview the delivered file to confirm that the sound and visuals suit the intended use.
Submit the video_edit operation, the full model ID, the video URL, and the editing instruction to POST /happyhorse/videos. With async, you can first obtain task_id and then query through /happyhorse/tasks; you can also provide callback_url to receive the result and obtain the video URL after successful completion.
The duration in video editing results represents the billable video duration, recorded as the combined length of the input and output videos, so it cannot be directly treated as the final video's playback length. The editing duration follows the source video, and actual usage is based on the statistics after task completion; see the pricing page for specific costs.
Choose a development approach or application suited to this model and follow its setup guide.
Python SDK Integration Guide
SDK installation, authentication and examplesTypeScript SDK Integration Guide
SDK installation, authentication and examples