16f4e93a3a
A ComfyUI companion that suggests core / available equivalents for custom nodes and replaces them in a workflow. - Backend (utfcn_core.py, __init__.py): read-only /utfcn/scan analysis that ranks equivalents in three tiers (curated → exact-signature → partial heuristic) from the live node registry. - Frontend (web/utfcn.js): on-add mode (Off / Suggest / Force auto-replace), bulk "Replace with core / available…" command + Extensions menu with a preview-then-confirm dialog, and a right-click single-node replace. The swap engine only rewires losslessly. - mappings.json: 7 hand-verified curated rules mined from installed packs (essentials, KJNodes, WAS); user_mappings.json for user overrides. - Docs + branding: README, icon and social banner (SVG + PNG), GPL-3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
107 lines
4.0 KiB
Markdown
107 lines
4.0 KiB
Markdown
<p align="center">
|
|
<img src="assets/social-preview.png" alt="UTFCN — Use The Core Nodes" width="100%">
|
|
</p>
|
|
|
|
# UTFCN — Use The F***ing Core Nodes
|
|
|
|
A ComfyUI companion that nudges your workflows back toward **core nodes**. Over
|
|
time a graph accumulates custom nodes that just re-implement things ComfyUI now
|
|
ships itself. UTFCN spots those and helps you swap them out — fewer dependencies,
|
|
more portable workflows.
|
|
|
|
It does three things:
|
|
|
|
1. **On add** — drop a custom node that has a core (or otherwise installed)
|
|
equivalent and, depending on the mode, UTFCN either shows a quiet tip or
|
|
(in **Force mode**) auto-replaces it with the equivalent on the spot.
|
|
2. **Replace across a workflow** — `Extensions ▸ UTFCN ▸ Replace custom nodes
|
|
with core / available…` (also in the command palette). It scans the open
|
|
graph and shows a **preview** of every swap before anything changes.
|
|
3. **Replace one node** — right-click any custom node ▸ **Replace with core /
|
|
available**.
|
|
|
|
Nothing is ever swapped without your say-so, and the engine only rewires slots
|
|
it can move *losslessly* — anything it can't map is reported, not guessed.
|
|
|
|
## How it decides what's equivalent
|
|
|
|
The backend reads the live node registry (real `INPUT_TYPES` / `RETURN_TYPES`
|
|
and each node's source module) and ranks candidates in three tiers:
|
|
|
|
| Tier | Meaning | Auto-applied? |
|
|
|------|---------|---------------|
|
|
| **curated** | a hand-written rule in `mappings.json` / `user_mappings.json` | yes (verified) |
|
|
| **exact** | identical input names+types and output types to a core/other-pack node | yes (verified) |
|
|
| **partial** | can structurally stand in (accepts all inputs, provides all outputs) but names/slots differ | suggestion only |
|
|
|
|
"Available" means core is preferred, and if there's no core match it will offer
|
|
an equivalent from a **different installed pack** as a fallback.
|
|
|
|
## Shipped equivalences
|
|
|
|
`mappings.json` ships a small, hand-verified set (each checked lossless against
|
|
the real core signatures):
|
|
|
|
- `GetImageSize+` (essentials) → core `GetImageSize`
|
|
- `MaskPreview+` (essentials) → core `MaskPreview`
|
|
- `BOOLConstant` / `INTConstant` / `FloatConstant` (KJNodes) → core
|
|
`PrimitiveBoolean` / `PrimitiveInt` / `PrimitiveFloat`
|
|
- `Convert Masks to Images` / `Mask Invert` (WAS) → core `MaskToImage` /
|
|
`InvertMask`
|
|
|
|
Rules for nodes you don't have installed are simply ignored. Everything else is
|
|
found live: exact-signature matches auto-apply, looser ones are suggested.
|
|
|
|
## Adding your own equivalences
|
|
|
|
To bless a partial match as safe, or to fix up slots whose names differ, add a
|
|
rule to **`user_mappings.json`** (merged over the shipped `mappings.json`,
|
|
survives updates):
|
|
|
|
```json
|
|
{
|
|
"rules": {
|
|
"SomeCustomNode": [
|
|
{
|
|
"to": "CoreNode",
|
|
"note": "why they're equivalent (shown in the preview)",
|
|
"inputs": { "old_input": "new_input" },
|
|
"widgets": { "old_widget": "new_widget" },
|
|
"outputs": { "old_output": "new_output" }
|
|
}
|
|
]
|
|
}
|
|
}
|
|
```
|
|
|
|
List targets in preference order (put the core node first). Any slot you don't
|
|
list is matched by identical name, then by type + order. After editing, run
|
|
`Extensions ▸ UTFCN ▸ Refresh equivalence index` (no restart needed).
|
|
|
|
## Settings
|
|
|
|
**UTFCN ▸ On add ▸ When adding a custom node that has a core / available
|
|
equivalent:**
|
|
|
|
- **Off** — do nothing.
|
|
- **Suggest** (default) — show a tip pointing at the equivalent.
|
|
- **Force (auto-replace with core)** — immediately swap it for the equivalent.
|
|
Force only ever applies **verified** matches (curated or exact-signature); it
|
|
never auto-applies a heuristic guess, and it never fires while you're opening
|
|
or importing a workflow — only on nodes *you* add. Undo with Ctrl+Z.
|
|
|
|
## Install
|
|
|
|
Clone into `ComfyUI/custom_nodes/` and restart ComfyUI:
|
|
|
|
```
|
|
git clone https://github.com/ethanfel/ComfyUI-UTFCN
|
|
```
|
|
|
|
No Python dependencies. The node adds a single read-only server route
|
|
(`/utfcn/scan`) and a frontend extension.
|
|
|
|
## License
|
|
|
|
GPL-3.0-or-later. See [LICENSE](LICENSE).
|