ComfyUI WAN Animate Developer Needed — Vast.ai Multi-GPU & Redis Queue Setup

# WAN Animate Reference-Video Motion Control — ComfyUI / Vast.ai Setup I need a developer to set up **WAN Animate reference-video motion control** in ComfyUI. The purpose is simple: use a reference video to control the movement of a supplied starting image and generate the final video. ## What I need - A tested ComfyUI workflow using WAN Animate for reference-video motion control. - Deployment on Vast.ai. - A multi-GPU worker setup, so additional Vast.ai GPU instances can process jobs in parallel. - A shared Redis queue. - A simple connection so my OpenClaw bot can submit jobs to Redis and receive job status and completed-video results. Each job will include: - Starting image - Reference motion video - Prompt and generation settings OpenClaw should only communicate with the Redis queue—not directly with individual GPUs or ComfyUI instances. The system should handle queued jobs reliably, process them using the WAN Animate workflow, and return the completed output location and status. Failed or interrupted jobs should be retried or clearly marked as failed. ## Cost priority My priority is low generation cost, targeting approximately **$0.01–$0.02 per second of finished video**. Please recommend the most cost-effective Vast.ai GPU and WAN Animate settings that can realistically work toward this target, stating any quality, resolution, or duration assumptions. ## Manual GPU scaling I need the system to be easy to scale manually based on my output needs. I should be able to add additional Vast.ai GPU workers when demand increases, and remove them when demand falls. Please provide a simple, documented process for: - Launching a new GPU worker from the prepared Vast.ai template or configuration - Connecting it securely to the shared Redis queue - Confirming it is ready to process jobs - Safely removing a GPU worker without losing jobs ## Deliverables - Tested ComfyUI WAN Animate reference-video motion-control workflow - Vast.ai deployment for one or more GPU workers - Redis queue and worker scripts - OpenClaw job-submission and result format - Multi-GPU parallel processing tested - Instructions for adding and removing GPU workers - Full workflow files, source code, configuration, and handover documentation ## Not required I do not need a dashboard, frontend, customer-facing UI, image generation, or image editing. Automatic GPU scaling is not required, but I must be able to add and remove GPUs myself easily. Please apply only if you have real experience with WAN Animate, ComfyUI, reference-video motion control, Redis queues, and Vast.ai. Include relevant examples and your recommended approach for the cost target. Please provide your hourly rate, estimated total hours, expected timeframe, and a maximum estimated cost for the full setup. State any assumptions or items that could affect the final estimate.