Anima Prompt Plus

The Anima Prompt Plus node (class AnimaPromptPlusClipEncode) is the "just make it work" version of Anima Prompt. It does everything the plain Anima Prompt node does - assemble quality + artist + character + clothing + pose + background + extra text in Anima's preferred order - and then encodes it for you. You hand it a CLIP, it hands you a ready-to-use positive CONDITIONING. No CLIPTextEncode in the middle. And as a bonus, it writes the resolved prompt into the image's PNG metadata so you can always recover exactly what you prompted.

The mechanism is straightforward under the hood: it composes the same prompt string as its sibling, then runs clip.encode_from_tokens_scheduled(clip.tokenize(text)) - the same path a normal CLIPTextEncode takes, just with the composed text pre-filled. If the clip input is missing or None, it raises a clear error telling you to connect a CLIP output from a checkpoint or text-encoder loader, which is friendlier than most.

Inputs and outputs

The inputs are Anima Prompt's full set, plus one:

• clip - the CLIP model to encode with. Feed it the CLIP output from your checkpoint loader (or a text-encoder loader).
• quality_prompt, artist_tags, character_tags, clothing_tags, pose_tags, background_tags, extra_prompt, separator - identical behavior to Anima Prompt, including the @artist normalization and the clean-join separator logic.

Outputs:

• positive - the CONDITIONING, wired straight into your KSampler's positive input.
• text - the resolved prompt string, if you also want it somewhere (a save node, a display, or a prompt bank).

That second output is genuinely handy: keep a text output around and you can log every prompt you generated without parsing PNG metadata.

Why the metadata matters

The node records the composed prompt under anima_prompt in the image's metadata. Two workflows later, when you find an image you love and want to know what produced it, the answer is right there in the file. The pack's author clearly built for people who iterate a lot and forget what they prompted - you'll be glad it exists the first time you come back to a good image with no memory of the prompt.

Install

Zero-dependency pack, no pip:

cd ComfyUI/custom_nodes
git clone https://github.com/nregret/Comfyui-Anima-Tools

Or ComfyUI Manager → "Anima Tools" → Install, then restart. Under AnimaArt, displayed as "Anima Prompt Plus".

If you're building a clean, repeatable Anima workflow, this is the node that collapses three steps into one: compose, encode, and archive the prompt with the image. The only time you'd prefer the plain Anima Prompt is when you want to inspect the string before encoding, or you're feeding a non-standard text encoder that the pack's composition would need to coexist with.