Building Short Midjourney V1 Videos with an Image-to-Video API

说明与用途

使用 Midjourney V1 通过 API 从单张图像生成短视频：准备图像、设置运动和时长、提交作业、轮询状态并下载 MP4。

问题

Generate short animated clips from a single image using Midjourney V1 while controlling motion, duration, and resolution programmatically.

结果

Successfully generated a 5‑second 480p MP4 from a single image with low motion, manual animation, and a single variant, demonstrating full programmatic control over Midjourney V1 video creation.

工作流

• 1. Prepare a publicly accessible image URL with clear subject and suitable aspect ratio.
• 2. Decide motion level (low/high) and animation mode (automatic/manual).
• 3. Choose video type (e.g., vid_1.1_i2v_480 for 480p SD).
• 4. Construct JSON payload with prompt, videoType, mode, animateMode, motion, and batch size.
• 5. POST to https://api.cometapi.com/mj/submit/video with Authorization header.
• 6. Capture returned job ID.
• 7. Poll https://api.cometapi.com/mj/status/{job_id} until status is "completed" or "failed".
• 8. On completion, retrieve video_url from result and download the MP4.

证据

Midjourney’s V1 video model is designed around a straightforward workflow: start with one still image and turn it into a short animated clip.

来源

The request accepts an image URL in prompt, a videoType such as vid_1.1_i2v_480, a generation mode, and an animateMode.

来源

The API is asynchronous, so the basic integration pattern is: Submit the generation request. Extract the returned job identifier. Poll the status endpoint. Download the video once the job reaches completed.

来源