Concurrent Submit | zhenzhen-creatify-aurora-fal

Creatify's Aurora is a talking-head / avatar video model: you give it a face image and an audio track, and it animates the face to the speech - the kind of job that content teams and dubbing pipelines run in bulk. The blocking zhenzhen-creatify-aurora-fal node does one at a time. This submit node is the parallel version: same fal-ai/creatify/aurora endpoint, same image-plus-audio contract, but each job goes into the shared video pool (10 workers) and returns a task handle instead of a finished clip.

If you've ever scripted ten lines for the same avatar, or generated ten voice tracks and needed them all animated, you know the serial pain. This is the node that turns that into one submit loop: image + audio pairs into the pool, collector hands back all ten videos in order.

How it works

Auto-generated concurrent wrapper, same as every submit node in the pack: the class copies the base node's inputs at import time, wraps its process() in a bounded ThreadPoolExecutor (video workers default to 10, set via COMFLY_VIDEO_CONCURRENCY), and returns a COMFLY_VIDEO_FUTURE. Collect with ComflyConcurrent_Video_Await ("Concurrent Collect Videos (10)"), which waits for the batch and returns VIDEO tensors in slot order.

Inputs mirror the base node: image_url and audio_url are the required fields (ignored when the corresponding image / audio inputs are connected), then api_key, prompt (optional - steers motion/expression), guidance_scale (1.0), audio_guidance_scale (2.0), resolution (480p default, 720p available), image_way, audio_way, poll settings, and skip_error. The seed here is the newer "execution seed" kind used for ComfyUI cache control - the tooltip is explicit that it isn't sent to APIs without a native seed parameter, so it's for re-run determinism in ComfyUI, not for varying output.

Installing it

One clone, whole pack:

cd ComfyUI/custom_nodes
git clone https://github.com/T8mars/Comfyui-zhenzhen

Restart ComfyUI, or install "Comfyui-zhenzhen" via ComfyUI Manager. No model downloads - Aurora runs hosted through the Zhenzhen proxy.

Common issues

Pack-level: account + credit required, 443 errors need VPN with TUN mode, first-run 500s are usually upstream - re-run. For this node: video pool is 10, not 30, and ten 720p avatar renders is ten paid calls plus ten long polls - do a single-clip sanity check at 480p before you fire the full batch. And audio_way defaults to upload, so if you're pointing at hosted audio instead, switch it to audio_url or the node will try to upload a URL string as a file.