CLIP Text Encode

Let's be honest about what this is: it's a stock CLIP Text Encode, the workhorse node that turns text into conditioning, wearing this pack's badge. ComfyUI core already ships one, so the reason to use this version isn't that it does something magical - it's that it lives inside the Smart Model Loader ecosystem, matches its PIPE-based conditioning contract, and has one deliberate design choice that trips people up.

That choice is in the node's own description: the text input is a forced connection, not a widget. You can't type a prompt into a little box on this node - you have to wire a text string into the text socket from a primitive or a prompt source. It's annoying the first time, and then you realize it's actually the whole point in a pipe-first workflow: the prompt becomes data flowing through the graph, coming from one authoritative source instead of being typed in five places. If you use this pack's Smart Model Loader and IO Checkpoint Loader, keeping the prompt on a wire keeps the whole pipeline consistent.

Inputs: clip (from the CLIP Loader, Model Loader, or Smart Model Loader's pipe) and text (the forced string input). Output: one conditioning socket feeding your positive and negative into the KSampler.

Under the hood it does exactly what ComfyUI's own encoder does: tokenize the text, then encode_from_tokens_scheduled - that "scheduled" bit is why you'd trust it in a video workflow, where conditioning can be spread across frames rather than applied as one static block. It's a thin, faithful wrapper, not a rewrite.

The one place it deviates from stock behavior is in error handling, and it's a genuinely good touch. If the clip input comes in as None - say you grabbed it from a checkpoint loader that has no valid text encoder - it raises the same explicit, readable error ComfyUI uses: "your checkpoint does not contain a valid clip or text encoder model." That message has rescued more beginners than any feature on the node.

Installing

cd ComfyUI/custom_nodes
git clone https://github.com/r-vage/ComfyUI_SmartModelLoader.git
cd ComfyUI_SmartModelLoader
python -m pip install -r requirements.txt

or ComfyUI Manager → ComfyUI Smart Model Loader, restart.

When you'd actually reach for it

• You're running this pack's pipe workflow and want one text source wired to everything.
• You need the scheduled encoding path for a video model and want it on a node you already trust.
• You keep hitting the stock node's empty-CLIP crash and want the version that tells you why in plain words.

Otherwise, honestly? The stock CLIP Text Encode is fine and you don't need to switch. This is a compatibility-and-convenience node, and knowing that is half the battle - its quirk is the only thing you need to remember about it.