Lucky Seed (7-Resonant)
A 'lucky' seed you can actually dial in — EddyLuckySeedNode explained
- seed
The name promises luck; the mechanism delivers determinism. EddyLuckySeedNode takes two numbers, does a bit of arithmetic, and hands you a seed you chose - the "lucky" part is entirely in your head, and that's fine. It's the manual, reproducible sibling to the same pack's random generator, and honestly it's the more useful half of the pair.
What it actually does
Here's the entire mechanism, verbatim from the source: output = int((n * 7) * resonance_factor), capped at 100,000. That's it - a formula, no randomness anywhere. Same n, same factor, same seed, forever. Where EddyIntSevenRandomV2 rolls the dice for you, this node is the dial: you pick an n, you get a seed, and you can reproduce that exact seed any time by plugging the same n back in.
Why bother? Two legit reasons. First, seed hunting: when you find a seed that gives you a great base image and you want to explore near it, a deterministic formula lets you nudge it by changing n by 1 and see a related variation - instead of pasting random 16-digit integers and praying. Second, sharing: small reproducible seeds travel well. "Use n=957" is a lot easier to pass around than a wall of digits, and your workflow stays fully reconstructible from the PNG metadata either way.
The inputs that matter
Just two, and both come with the author's own tooltips:
n(INT, default 957, range 1–12173) - the seed dial. Default 957 spits outint(957*7*1.15)= 7703.resonance_factor(FLOAT, default 1.15, range 1.0–2.0, step 0.01) - despite the mystical name, this is just a multiplier. Crank it toward 2.0 and you reach the high end of the node's tiny 100,000-cap output range (maxnthat stays under the cap shrinks accordingly).
The single output, seed (INT), wires straight into your KSampler's seed port - or into EddyIntSevenRandomV2's seed input, if you want to chain the numerology into the random node. Note the cap: this node can never produce a seed above 100,000. That's a deliberate design constraint from the author, and it means the "resonance" branding is doing a lot of heavy lifting - there's no evidence seeds in this range generate better than any other, and the honest take is that this is a deterministic small-seed picker with a fun name.
Installing it
Same pack, same story as its sibling: pure Python stdlib, zero dependencies, no requirements.txt, no model downloads. The README is one ASCII diagram and an image - no install instructions, because there's nothing to install beyond the folder itself.
- ComfyUI Manager: search "Seven Resonance" (or "Eddy") and install
ComfyUI-EddySevenResonance, then restart. - Manual:
then restart ComfyUI. No pip, no weights, nothing else.cd ComfyUI/custom_nodes git clone https://github.com/eddyhhlure1Eddy/ComfyUI-EddySevenResonance
Gotchas
This pack is so new and anonymous that it has essentially zero community presence - I checked, and nothing anywhere mentions it - so there's no war story to report. What the source does tell you:
- Small-seed ceiling. 100,000 max means you can't reach the high seed space at all. If a sampler's output is sensitive to seed magnitude in your workflow, this node narrows your options by design.
- No luck involved. If you genuinely want variety, use the sibling random node or just let the stock seed widget randomize. This node's whole point is a seed you can hand-pick and reproduce - lean into that and it's quietly handy.
Would I reach for it? If I wanted deterministic, dialable, shareable seeds, yes - it's the cleanest 30-line implementation of that idea I've seen. If I expected it to make my images luckier, no. It doesn't, and it never claimed to in the code. The luck was you all along.
Inputs (2)
| Name | Type | Default | Description |
|---|---|---|---|
| n | INT | 9571–12173 | n value, output = int((n*7)*resonance) |
| resonance_factor | FLOAT | 1.151–2 | Resonance multiplier (default 1.15) |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| seed | INT | — |