Seedream 5.0 Lite · Text to Image
Pinned 2K sizes, no surprises
- images
- task id
Three widgets: prompt, size, output format. No sampler, no seed, no resolution tier guessing. Seedream 5.0 Lite Text to Image is the least dramatic node in this pack and the one most likely to end up in a batch pipeline, because it does the one thing batch work needs - it always gives you the exact size you asked for, straight off a fixed list, at 5.18 credits (~$0.052) an image.
If you've ever had a hosted image model hand back "roughly" your aspect ratio and then spent an hour resampling outputs to make them line up, this is the fix.
What's happening under it
The node calls bytedance/seedream-5.0-lite/text-to-image on PopNow's servers, through the pack's shared plumbing: quote with jobs/quote first (so the price is known before anything is held), submit with an idempotency key, poll, download, decode into a ComfyUI IMAGE batch. Nothing runs locally and no weights exist to download - Seedream is closed, so the API call is the only route.
Practical effect on your machine: no VRAM used, no cold start, and the node works on a laptop, a Mac, or a CPU-only box. The cost is per image and it's real, so keep the arithmetic in your head: a 40-frame storyboard is 207 credits.
The three inputs
prompt- subject, composition, style, lighting. Lite handles natural description and legible in-image text well; quote anything that must render literally.size-width*height, fixed list:2048*2048(default),2304*1728/1728*2304,2848*1600/1600*2848,2496*1664/1664*2496,3136*1344. Square, 4:3 and 3:4, 16:10-ish wide, ultrawide. That's the whole menu - no 4096 sizes here, that's 4.0 and Pro territory.output_format-jpeg(default) orpng. PNG if anything downstream cares about compression.
Outputs
images - a batch - and task id. Standard wiring: Save Image for a look-see, or into your local detailer/upscaler if you're using Lite as the cheap draft layer.
No alpha output on this node, and no background switch. If you need a cutout, the Flash Edit node in the same pack is the one that offers transparency - Lite doesn't, and it's easy to waste ten minutes looking for a setting that isn't there.
Install
cd ComfyUI/custom_nodes
git clone https://github.com/popnow-ai/popnow-comfyui
Or use ComfyUI Manager: search PopNow, Install, restart. Then add a key from popnow.ai/api/console/keys in Settings > PopNow, or set POPNOW_API_KEY in ComfyUI's environment, and restart once more - the node list is generated from your account's live catalog at startup. No dependencies to install; the pack declares an empty list on purpose.
Where people trip
PopNow 40201- not enough credits. Unspectacular per image, but you'll hit it mid-batch if you weren't counting.PopNow 400naming a field - asizethat isn't on the list. It's a dropdown; you can't type1920*1080.PopNow 40004- the model won't serve that combination; change the field the message names.PopNow 40310- email not verified. It does block runs, and the message is unhelpfully terse.generation_failed- refunded. Rerun once before assuming your prompt is the problem.- Cancelling stops your wait, not the job. The run finishes on PopNow, bills if it succeeds, and the result is in your API console logs.
- A model visible on popnow.ai with no node in ComfyUI - restart; the catalog is read at startup.
Is Lite worth 20% more than Flash?
Flash is 4.35 credits and offers three resolution tiers plus transparency; Lite is fixed-size, no alpha, and about a fifth more expensive. On paper that reads backwards. The reason to pick Lite is that its output sizes are pinned to a specific set of 2K frames that hold a series together, and in practice its look sits between Flash's quickness and Pro's sharpness. If you're generating one image, use Flash. If you're generating forty that have to match, Lite's rigidity stops being a constraint and starts being the feature.
One more honest note: a custom node that spends money per click is a different category from a sampler, and this whole pack is that. Read what's on the node before you loop it - the price badge updates as you change settings, which is a small but genuinely good bit of design.
Inputs (3)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | Describe the image to generate: subject, composition, style, and lighting. | |
| size | COMBO | 2048*2048 | Output image size in pixels, written as width*height. |
| output_format | COMBO | jpeg | File format of the returned image. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| images | IMAGE | — |
| task id | STRING | The PopNow task id, for the API console logs or support. |