Concurrent Submit | Zhenzhen MJ Video Extend
Extend Midjourney videos 4 seconds at a time — and queue the extensions
- task
Midjourney video clips come out short, and when you need more, you extend them - a few seconds at a time, chained off the previous result. Comfly_mj_video_extend is the pack's node for exactly that, and the author's README is admirably plain about how it works: you give it the task ID of a previously generated video and an index 0–3 for which of the four output videos to extend, and it extends that clip by about four seconds, up to four extensions per clip. This Concurrent Submit twin adds the queue: run a batch of extensions in the 10-worker video pool instead of chaining them one painful round-trip at a time.
The inputs that matter
- task_id - required. The task ID of the video you're extending. Must be a previous
Comfly_mj_video(or prior extend) output - this is a pure follow-on operation. - index - required. 0–3, selecting which of the four videos from that task to extend. If your task returned four videos,
indexis how you say "extend the second one." - api_key - your Zhenzhen key.
- seed / skip_error - cache-control seed and the fail-silently toggle.
That's it. No prompt, no motion - this node doesn't re-imagine, it just keeps the clip going.
The concurrent flow
Output is a task (COMFLY_VIDEO_FUTURE), not a video. Wire it into ComflyConcurrent_Video_Await (10 slots) with your other jobs and the collector returns the extended clips in slot order with a status JSON. Since extension is inherently serial (extend → extend again), the batch pattern works best when you're extending different clips in parallel - one node per clip - rather than trying to chain one clip through concurrent jobs, which defeats the ordering.
Install and gotchas
Standard pack install: T8mars/Comfyui-zhenzhen via ComfyUI Manager or git clone https://github.com/T8mars/Comfyui-zhenzhen into custom_nodes, then restart. No local models, standard deps, paid token, overseas servers (proxy from China).
The mechanics to keep straight, straight from the changelog: each extension adds ~4 seconds, you can extend a clip at most 4 times, and it's billed per extension. So a clip going from short to long is multiple billable calls on the same task ID - keep the task_id and index wiring honest, because extending the wrong one of the four videos is a very Midjourney way to waste a call.
Inputs (5)
| Name | Type | Default | Description |
|---|---|---|---|
| task_id | STRING | — | |
| index | COMBO | 0 | 4 options: 0, 1, 2, 3 |
| api_keyopt | STRING | — | |
| skip_erroropt | BOOLEAN | false | 开启后,节点失败时不报错、按旧行为返回默认空结果;关闭时(默认)失败直接抛出错误。 |
| seedopt | INT | 00–18446744073709550000 | Execution seed for ComfyUI cache control. Fixed reuses the cached result; randomize/increment/decrement requests a new run. This compatibility seed is not sent to APIs that do not expose a native seed parameter. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| task | COMFLY_VIDEO_FUTURE | — |