A 90-second 2D-rendered walkthrough of the deterministic core, the workflow, and the cinematic-output envelope. Watermarked. Downloadable.
The Studio demo above proves the deterministic core. Once you export a bundle, you point it at a video diffusion runtime that runs on your own GPU — see the BYOC starter pack for the full pipeline. Three options below, in increasing order of polish.
If you already use ComfyUI: download our workflow file, import it via the Load button, swap in your local video model checkpoint, paste in the prompt and seed from your exported bundle, queue. Render lands in your output/ folder.
If you have diffusers + torch with CUDA: download our slim Python runner, edit the render_with_diffusers() function to call your local model, then python neyralfoundry_runner.py bundle.json. Watermark applied automatically.
One-command CLI that bundles model + watermark: neyralfoundry render bundle.json. Ships in beta. No model setup, no glue code, no Python knowledge required.