Concurrent Submit | Zhenzhen Google Veo3

This is the flagship Google video node in the Comfyui-zhenzhen pack, and it's a twelve-model menu in one box: veo3, veo3-fast, veo3-pro, the -frames image-to-video variants, the 3.1 line (veo3.1, veo3.1-fast, veo3.1-pro, veo3.1-components), and the 4K pair (veo3.1-4k, veo3.1-pro-4k, veo3.1-components-4k). Veo's signature is native audio - dialogue and sound effects baked into the video rather than bolted on - and per the KB's closed-source-models essay that's still a genuine capability gap for open video. If you want that, this is how it enters a ComfyUI graph.

The Concurrent Submit wrapper matters here more than anywhere: 4K Veo jobs are the slowest, most expensive thing in this pack, and running them one-at-a-time is a treadmill. Batch them.

How the concurrent bit works

The plain Comfly_Googel_Veo3 node blocks until the clip lands. The Submit variant runs it in the pack's shared video pool (10 workers) and returns a task immediately. Collect via Concurrent Collect Videos (10): wire task_1 … and the collector waits for the batch, returning videos in slot order plus a status string. failure_mode = fail-fast (default) or placeholder.

The inputs that matter

• prompt (required) - the shot description.
• model - the twelve-model pick list. Fast = cheap and rough, Pro = premium, -frames = image-to-video (connect image1), -components = multi-image reference (up to 2), and anything with 4k costs accordingly. The README's pricing table is the honest spec sheet - Veo prices changed a lot over the pack's life, so check the shop before picking Pro 4K.
• enhance_prompt - default false. Flip on to auto-translate Chinese prompts to English (Veo's native language).
• aspect_ratio - 16:9 or 9:16. Veo's other ratios live in the bigger family nodes; this one keeps it to the two.
• image1 … image3 - reference images for the frames/component models.
• enable_upsample - default false; upscaled output costs more.

Output: one task (COMFLY_VIDEO_FUTURE) → Collector → videos.

Installing it

Ships in Comfyui-zhenzhen:

• ComfyUI Manager → Install Custom Nodes → search "Comfyui-zhenzhen" → install → restart.
• Or cd ComfyUI/custom_nodes && git clone https://github.com/T8mars/Comfyui-zhenzhen and restart.

No model downloads - pure API. You need a paid token from the author's storefront (ai.t8star.org), in api_key.

Gotchas

The README's Veo section is basically a hazard report: models get pulled for days at a time (veo3.1-fast was temporarily suspended once), token groups matter (the spare "veo&grok" groups are where the cheap Veo lives), and 4K runs are minutes long with real failure odds - when ComfyUI times out, the author's advice is to check the shop's async-task page for the task ID before rerunning and re-spending. Also: Veo automatically generates audio and you can't turn it off (per the README). And the seed input here is ComfyUI-cache-only, not a real Veo seed - "Fixed" reuses the cached clip rather than re-rolling.