Text Parse A1111 Embeddings
Port Automatic1111-style embedding references into ComfyUI syntax
If you've ever copied a prompt over from Automatic1111 or a similar UI, you'll have run into this: A1111-style prompts reference textual-inversion embeddings by bare filename - just the name, no special syntax - while ComfyUI wants them wrapped as embedding:filename. Paste an A1111 prompt straight into ComfyUI and any embedding references in it just get read as plain words instead of triggering the actual embedding. Text Parse A1111 Embeddings exists specifically to fix that mismatch.
Per the README: it converts "embeddings filenames in your prompts to embedding:[filename]] format based on your /ComfyUI/models/embeddings/ files." It scans your actual embeddings folder to know what names to look for, then rewrites matching filenames in your prompt text into the syntax ComfyUI's CLIP text encoder actually understands.
How it works
Feed it a prompt as text, and it checks the words against the filenames present in your ComfyUI/models/embeddings/ directory. Anything that matches gets rewritten into ComfyUI's embedding syntax; everything else in the prompt passes through unchanged. Because it reads your actual embeddings folder rather than a fixed list, the set of names it recognizes is whatever you've actually got installed - add a new embedding file and it becomes something this node can pick up.
Where it's genuinely useful
Porting prompts from Automatic1111, Forge, or any other UI that uses the bare-filename embedding convention, without manually hunting through the prompt for every embedding reference and rewriting it by hand. It's a one-time conversion step you drop in front of your CLIP Text Encode, not something that needs to run every generation once the prompt's already in ComfyUI's native format.
What it takes and gives back
A STRING of prompt text goes in, and a STRING with embedding references rewritten into embedding:filename syntax comes out - plug that straight into your text encoder node in place of the raw prompt text.
How to install it
ComfyUI Manager: search "WAS Node Suite," install, restart. Manual:
cd ComfyUI/custom_nodes
git clone https://github.com/WASasquatch/was-node-suite-comfyui
then, from inside that folder, install requirements - path/to/ComfyUI/python_embeded/python.exe -s -m pip install -r requirements.txt on portable, pip install -r requirements.txt on system Python. Restart ComfyUI; it's under WAS Suite.
Common issues & troubleshooting
If an embedding isn't getting converted, the most likely cause is a filename mismatch - the node matches against what's actually in ComfyUI/models/embeddings/, so a typo in the prompt, a different file extension than expected, or the embedding simply not being in that folder yet will all cause it to pass through as plain text instead of getting rewritten. Confirm the file is actually present in that folder before assuming the node is broken.
This is a text-parsing node with no external dependency of its own, so it's unlikely to fail in isolation. The pack-wide risk still applies: WAS Node Suite has had no active development since December 2023, and the recurring complaint since then is the entire suite going "Import Failed" after a ComfyUI update - a shared-dependency version clash, not this node specifically. If it happens, reinstall requirements with your embedded Python interpreter, not a system one.
Inputs (0)
No inputs
Outputs (0)
No outputs