Loading...
Loading...
Numonic captures every ComfyUI workflow from every artist on your team, automatically. Node graphs, prompts, LoRAs, samplers, seeds. All searchable, all reproducible, all there when the client asks for it months later.
One artist's setup is intricate enough. Your studio has three or five or eight artists each running their own ComfyUI, each saving workflows to their own machines, each making the same custom nodes break in different ways. When a client comes back to a previous look, no one knows which artist made it, on which version of which node graph, with which checkpoint. Numonic captures every artist's ComfyUI output the moment it lands on disk.
Connect each artist's ComfyUI output folder once from the browser. Numonic watches the folder, captures every PNG and MP4 along with the embedded workflow JSON, and adds the asset to your studio's library. The artist's save habits do not change. The studio's memory just starts working.
Numonic reads the workflow embedded in each generation and keeps the whole graph: checkpoints, LoRAs and VAEs with hashes, sampler, scheduler, steps, CFG, seed, dimensions, and custom node configs. Everything you need to reproduce the work, captured automatically.
Search your studio's entire ComfyUI output by any parameter, by checkpoint, by LoRA, by seed, or by visual similarity.
seed:88934lora:3cfg:8One click on any captured asset reopens the original workflow directly in ComfyUI Cloud. Re-run it, tweak it, no version mismatches.

Every model and LoRA detected from your workflows lands in a searchable library, with usage tracked across artists.
Two options. Connect your ComfyUI output folder from your browser and new images upload automatically, with no app install. Or drag and drop files in the web interface. Either way, no workflow changes are required.
No. Numonic runs completely separate from ComfyUI. Connected Folders monitors your output folder and uploads files after they are saved, with zero impact on generation speed.
Everything embedded in your files: node configurations, prompts, model names, LoRA settings, sampler parameters, seed values, and custom node settings. Enough to fully reproduce any generation.
Yes. The metadata stripper removes prompts, seeds, model names, and LoRA settings before sharing, so you keep your techniques private while optionally adding EU AI Act disclosure metadata.
Yes. 5 GB storage, unlimited workflow captures, full search, and workflow export. Paid tiers add team collaboration and higher limits.
Yes. MP4, MOV, WebM, and MKV files get full metadata extraction, including workflows from MP4 comment atoms, plus codec, duration, and framerate data.
Start free for personal use. Move to the Studio tier when you need shared capture across artists.
Start freeNo credit card to begin