Drop a ComfyUI PNG here
or click to choose · or paste (Ctrl/Cmd + V)
processed in your browser · never uploaded
Drop or click to replace
No ComfyUI workflow found in this image.
Summary
- Nodes
- —
- Models
- LoRA
A workflow extractor for when you find a great ComfyUI image online and want to rebuild it in your own setup. ComfyUI embeds two JSON blobs in the metadata (tEXt/zTXt/iTXt chunks) of every PNG it generates: the workflow (the UI node graph) and the prompt (the API-format execution). Drop or paste a PNG and this tool parses those bytes, then summarizes the node count, the list of node types, and the checkpoints (models) and LoRA in use. You can copy the workflow JSON or save it as a .json file, and loading it into ComfyUI's "Load" restores the entire node graph. This is a read-only tool — it never edits or alters the image. Nothing is sent to any external API or server; all parsing happens locally in your browser, so you can safely inspect other people's work or your own unpublished images. Note that PNGs without this metadata — files converted to JPEG/WebP, or images re-compressed by social platforms — may not yield any JSON. All output is rendered as plain text, so metadata strings are never executed as HTML.
How to use
- Drag and drop a PNG that ComfyUI produced onto the drop zone, click to choose a file, or paste it from the clipboard.
- Review the summary (node count, node types, models / LoRA used) and the workflow (UI graph) / prompt (API) JSON.
- Copy the JSON or save it as .json, then load it into ComfyUI's "Load" to restore the node graph.
FAQ
Is my image uploaded to a server?
No. All PNG metadata parsing happens inside your browser, with nothing sent to any external API or server. The image you load is never uploaded or stored — it is processed only on your device, so you can safely inspect other people's work or unpublished images.
What's the difference between the workflow and the prompt?
The workflow is ComfyUI's UI node graph (wiring and layout included); loading it into ComfyUI's "Load" restores the whole canvas. The prompt is the API-format JSON of what actually ran, showing node input values (model, prompt, sampler settings, and so on). This tool extracts and shows both.
Can I extract a workflow from a JPEG or WebP?
Usually not. ComfyUI workflows live in PNG text chunks, so converting to JPEG or WebP typically strips them. Use the original PNG — the file ComfyUI saved directly.
Can I get the workflow from an image posted on social media?
Often no. X (Twitter) and most platforms re-compress images on upload and strip metadata, so a downloaded image usually has nothing left to extract. You'd need the original PNG or workflow JSON shared separately by the author.
What are the node-type and model/LoRA summaries good for?
They let you see which checkpoint (model) and LoRA were used and which nodes (upscalers, ControlNet, etc.) are involved — without reading the whole JSON. It's a quick checklist of what you need to gather before reproducing the result.
Can I use the extracted JSON directly in ComfyUI?
The workflow JSON loads straight into ComfyUI's "Load." But it will error if the models, LoRA, or custom nodes it references aren't installed in your setup. Use the summary to check the required models and LoRA, install what's missing, then load it.