Anima Prompt Studio Advanced v2

The original Anima Prompt Studio Advanced spat out a wall of ten string outputs - positive, negative, mod-guidance quality tags, metadata, width, height, two booleans. It worked, but every downstream node had to accept that exact shape. Prompt Studio Advanced v2 is the pack's response to its own success: same editor, one output socket. It emits a single EASYUSE_ANIMA_PROMPT_DATA dict, and the whole new AiO flow is built around it.

What it is

This is the editor you'll actually live in. It keeps the reorderable positive/negative field groups from the Advanced node - quality, artist, trigger, general, and a NAIA field - plus resolution buckets, wildcard control, Mod Guidance routing, and artist-mix settings, all driven from popups so the node body stays small. The change is downstream: instead of ten sockets, one dict socket that downstream nodes read by key.

That dict is the pack's lingua franca now. It stores your prompt fields, the resolved width/height, wildcard-expanded text, Mod Guidance flags, and the artist-mix configuration under named keys. It also carries the old ten values under a compatibility key and mirrors every required v2 input under parameters, so nothing breaks when fields get added later - downstream nodes are told to read keys, not output-slot order.

The inputs that matter

Most of the required inputs are internal JSON payloads you shouldn't hand-edit (advanced_fields, resolution_bucket, wildcard_seed, and friends). The ones with real meaning:

• use_naia - the front-end turns this into the "Fill from NAIA" button on the NAIA field, pulling a fresh random prompt from your local NAIA endpoint.
• consume_naia_on_queue - one-shot mode. A successful NAIA fill is saved with the request flag off, so a loaded workflow reuses the stored prompt instead of pinging NAIA again. Leave it on.
• use_anima_mod_guidance and use_negative_anima_mod_guidance - route quality fields out of the main prompt and into the Mod Guidance path (which the AiO Generator applies as a model patch via ComfyUI-Spectrum-KSampler).
• artist_mix_mode - when it's not off, artist-field text is pulled out of the base prompt and handed to Prompt Data Conditioning to build mixed conditioning branches.

The output is one socket: EASYUSE_ANIMA_PROMPT_DATA. Connect it to Easy Use Anima Input, and you're done - the whole pipeline reads from the dict.

A detail people miss

Wildcards expand at queue time, and if a wildcard in a v2 positive field resolves to a <lora:name:weight> tag, that tag becomes structured LoRA data inside the prompt dict. A connected AiO Generator then applies it automatically after your explicit LoRA stack. Same for A1111/LoraManager-style tags - v2 recognizes them, keeps them structured, and the generator applies them. That's the "prompt embeds its own LoRAs" trick done properly.

Install and gotchas

It's in the pack, so:

cd ComfyUI/custom_nodes
git clone https://github.com/n0va39/ComfyUI-EasyUseAnima
cd ComfyUI-EasyUseAnima && pip install -r requirements.txt

then restart ComfyUI (Manager: search "ComfyUI EasyUse Anima").

Gotchas: the hidden JSON inputs are workflow-schema compatibility - edit them by hand and you'll likely corrupt a workflow. If you open an old workflow that expects the ten string outputs, don't panic: the v1 Advanced node still exists, and the EASYUSE_ANIMA_PROMPT_DATA node (the unpacker) turns a v2 dict back into the old outputs if you need to bridge legacy wiring. If nothing shows up in the NAIA field, that's the NAIA service or the global settings (localhost by default), not this node.