🔶 Load Checkpoint +Prompt

Strip away the frame logic that most of this pack revolves around, and you get this: a plain checkpoint loader with your positive and negative prompts baked into the same node, plus a couple of "add-on" text fields for stacking extra tags on top without editing your main prompt. If you're not doing the frame-by-frame animation thing at all and just want fewer nodes cluttering a normal txt2img graph, this is the one to reach for - it's functionally a CheckpointLoaderSimple plus two CLIPTextEncode nodes collapsed into one.

The "add" fields are the actual reason to pick this over the stock loader. Say you've got a base prompt you reuse across a project, but want to bolt on "cinematic lighting, 8k" for one render without retyping the whole thing - add_positiv_txt gets appended rather than replacing, so your core prompt stays untouched and reusable.

Inputs and outputs

• Checkpoint - required, your model dropdown, same as any standard loader.
• Positiv, Negativ - required, multiline text, your main prompt and negative prompt.
• add_positiv_txt, add_negativ_txt (optional, multiline) - extra text appended on top of the main prompt fields, for the "same base prompt, small variation" use case.
• add_lora (optional, LORA type) - wire in a LoRA from one of Chaosaiart's LoRA nodes to fold into this load.
• Outputs: Info (a string report - useful if you're using pythongosssss's show-text node the README credits, to confirm what actually got loaded), MODEL, POSITIV, NEGATIV, VAE - wire straight into your KSampler.

Installing it

ComfyUI Manager: Install via Git URL → https://github.com/chaosaiart/Chaosaiart-Nodes. That's the install path the README itself documents. A title search for "Chaosaiart" in the Manager should also find it.

Manual:

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

Linux (venv activated first, if you use one):

pip install opencv-python
pip install tqdm

Windows: run the bundled Install_windows script instead. Restart ComfyUI either way. No models to fetch - it loads whatever checkpoints and LoRAs you already have on disk.

Common issues

Don't expect this to do anything the frame nodes do. There's no activ_frame input here - it's the plain, always-the-same-thing-every-run loader. If you came looking for the checkpoint-swapping behavior, that's chaosaiart_CheckpointLoader (the "by Frame" version) or the CheckpointPrompt_Frame + FrameMixer combo, not this one.

VAE comes from the checkpoint, not a separate loader. Like most all-in-one loader nodes, if your checkpoint doesn't bundle a good VAE (common on older SD 1.5 merges), you'll get the classic grey, washed-out output - the fix is the same as with any checkpoint: source a matching standalone VAE and load it separately rather than relying on this node's built-in one.

One shared module for the whole pack. All Chaosaiart nodes live in a single Python module - if opencv-python or tqdm failed during install, the entire pack (not just this node) won't show up. Check ComfyUI's startup log for an import error before troubleshooting further.