Anima Random Artist Selector

If you've spent any time with Anima, you've heard the advice: the model lives or dies by its artist tags. "Artist tags is pretty overpowered on this model" is basically the community consensus, and the recommended onboarding is to find a dozen styles you like and prompt from those. AnimaRandomArtistSelector is a tiny built-in lottery for exactly that - pick a curated list, hit go, get a seeded random artist (or up to five) to prompt with.

What it does

It's three inputs and one string out:

• mode - Core Artist or Recommend Artist, two different built-in lists
• count - 1 to 5, how many artists to draw (no duplicates)
• seed - the RNG seed, so the same seed always hands back the same artists

The output, artist_text, is a comma-joined string like asanagi, michiking, abubu - ready to drop straight into your Anima positive prompt, ideally through the pack's Anima Artist Formatter if you want them in (@name:weight) syntax.

The two lists live in the pack repo: Artist_Core.txt has 42 hand-picked names, Artist_Recommend.txt a broader 240. The "Core" list is the safe starter set; "Recommend" is where you go once you've exhausted the first 42 or want more range.

How it works

Under the hood it's refreshingly direct: seed the Python RNG with your seed, read the chosen file line by line into a list, then random.sample() picks count entries with no repeats. No network, no models, no fuzzy matching - the artists are whatever the author curated into those text files.

That seeded-RNG detail is the part worth caring about. Because the selection is seeded, a given seed is reproducible: rerun the workflow with the same seed and you get the same artists, which is what you want when you're hunting for the seed that produced a good image. Change the seed and you roll fresh. If you want the node to pick a different set every run, feed it a varying seed (the pack's Seed Generator node, or a per-run random source, works).

Where it fits

This is a discovery and variation tool, not a precision instrument. The honest framing: the curated lists are a starting point, and the Anima ecosystem has much deeper artist exploration tools (animadex, the Anima Style Explorer with its 20,000+ artist database) if you want to go hunting deliberately. This node is for the lazy-mode workflow - generate, get surprised by a style, note the seed, iterate. It also slots nicely into a batch loop: vary the seed, get a different artist each frame, and force yourself to actually explore styles you wouldn't have picked manually.

One workflow note: artist_text is a plain comma string, so it wires into any prompt concatenation or the formatter node. If you're also using wildcard files, this node is a decent complement - it covers the artist dimension with a curated list while your wildcards handle the rest.

Install

It's part of the ComfyUI Seed Wildcard Pack:

cd ComfyUI/custom_nodes
git clone https://github.com/sinanzoo2nd/ComfyUI-Seed-Wildcard-Pack

or ComfyUI Manager → search Seed Wildcard → install, then restart. No dependencies beyond the Python stdlib, no model downloads, no requirements.txt to satisfy. The artist lists ship inside the repo, so there's nothing extra to fetch.

Gotchas

• If the list file ever goes missing or reads empty, the node returns the string None instead of artist names - that's the designed fallback, not a crash, and it's a hint that the pack folder got moved or partially deleted.
• count is capped at 5 by the UI, and random.sample can't draw more names than the list holds - with 42 core artists you can't ask for more than that in a single call (and shouldn't).
• The lists are the author's taste, not the model's full knowledge. Anima knows far more artists than these 282 names; treat this as a curated shortcut, not a ceiling.