Random Entry

Random Entry reaches into a list and pulls out one item at random. It's the node behind "surprise me" workflows - random prompts from a curated list, random reference images, random anything - and the seed input means you can have that randomness be a total coin flip or a perfectly repeatable pick.

Why you'd reach for it

Most ComfyUI workflows are deterministic on purpose, and that's exactly why a random node is useful: it's the one deliberate source of chaos. Feed it a list of prompts from String To List and it picks one to feed your KSampler, so a batch run explores different prompts without you re-typing anything. The seed gives you the best of both worlds - lock a seed to reproduce an exact pick, or set -1 to roll fresh every run.

How it works

The node takes its seed and calls random.seed() with it, then random.choice() on the list. Two details matter:

• seed -1 (the default) means "don't seed at all" - Python seeds from time/entropy, so every execution picks something different. The node also signals this to ComfyUI's caching by reporting "always changed", forcing a re-run each time.
• any other seed makes the pick deterministic: same seed, same list, same entry, every time.

The node is declared INPUT_IS_LIST, so the input is a real list. Empty list? You get an empty string back rather than a crash.

The inputs that matter

• seed - -1 for true randomness, anything else for reproducible picks.
• string_list - the list to choose from. Needs to come from a list-producing node.

One output:

• Entry - the chosen item, as a STRING.

Installing it

Pack install: ComfyUI Manager → Install Custom Nodes → search "ComfyUI-TinyBee", install, restart. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/TinyBeeman/ComfyUI-TinyBee

It's under 🐝TinyBee/Lists. Stdlib only, no dependency fuss.

Common issues

The biggest trap is feeding a single string instead of a list - you'll get that one string back every time and wonder why "random" never changes. Also note that determinism depends on the order of the list: the same seed picks a random index, so if the list order changes, the pick changes. And if you're hunting for a bug, remember seed -1 is intentionally non-reproducible; if you need to compare runs, pin a real seed and use Iterate Seed to step through controlled variations instead.