🔶 Frame_Prompt Clip Endcode

The name's typo ("Endcode" for "Encode") is a small, honest tell of what this pack is: a solo dev's tool built to solve their own animation workflow, not a polished commercial release. Function-wise, it's straightforward once you know what it's for - it's Chaosaiart's own stand-in for a standard CLIPTextEncode, except instead of taking plain text it takes a FRAME_PROMPT bundle and encodes that against your model and CLIP, handing back a MODEL plus encoded POSITIV/NEGATIV conditioning ready for your KSampler.

Worth being upfront about a real gap here: FRAME_PROMPT isn't produced by any of the nodes covered elsewhere in this batch - it comes from another node further up the pack's prompt-by-frame chain that isn't part of this set. If you land on this node without already having a FRAME_PROMPT source in your graph, you'll need to track down whichever Chaosaiart prompt node emits that type (check the pack's GitHub workflow examples for the exact wiring) before this one does anything useful.

Inputs and outputs

• model - required MODEL, your loaded checkpoint's model output.
• clip - required CLIP, same source as any standard text encode.
• frame_prompt - required, FRAME_PROMPT type - a frame-scoped prompt bundle from elsewhere in the pack. This is the input that makes the node specific to Chaosaiart's ecosystem rather than a drop-in replacement for the stock encoder.
• Outputs: MODEL (passed through, wire onward to your sampler setup), POSITIV, NEGATIV - the encoded conditioning.

Like the other custom types in this pack, FRAME_PROMPT only connects to other Chaosaiart nodes - you can't feed it a plain string or a standard conditioning object.

Installing it

ComfyUI Manager: Install via Git URL → https://github.com/chaosaiart/Chaosaiart-Nodes, exactly as the README documents. A title search for "Chaosaiart" should also find it.

Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/chaosaiart/Chaosaiart-Nodes

Linux (activate your venv first, if you use one):

pip install opencv-python
pip install tqdm

Windows: run the Install_windows script bundled in the folder. Restart ComfyUI afterward. No extra downloads needed for this node - it encodes text against a model and CLIP you've already loaded.

Common issues

Nothing to plug into frame_prompt. If you don't already have a FRAME_PROMPT-producing node in your graph, this node has nothing to do - it's not a general-purpose text encoder, and there's no fallback to plain text. Grab one of the pack's example workflows from GitHub to see the intended upstream wiring rather than guessing.

Expecting it to behave exactly like a stock CLIPTextEncode. It shares the same job at the output end (producing usable conditioning), but the input side is entirely Chaosaiart's own typed system - this isn't a swap-in replacement you can drop into a non-Chaosaiart graph.

Whole-pack import failures. Every node in this pack, including this one, lives in a single Python module - a broken opencv-python/tqdm install takes all ~20 nodes offline at once. Check the ComfyUI startup console for an import error first.