NS WaveSpeed Nano Banana Pro Text to Image Ultra
8K text-to-image, from a prompt alone
- client
- output_image
An 8K image generated from a text prompt, with zero local GPU. That's the entire pitch of Nano Banana Pro Text to Image Ultra, and it's worth pausing on, because "8K native output from a generator" is not something the local ecosystem does - even 4K usually means an upscale pass or a tiling project.
It's part of the Symbiotica pack's neuralsins/WaveSpeed family. This is the top rung of the Pro text-to-image ladder: the single-output Pro T2I node stops at 4K, this one adds an 8k tier. The tooltip quotes the meter at $0.15 for 4k and $0.18 for 8k - which is cheap for what you're getting, and notably cheaper than the Pro Edit node's $0.24 4k tier.
How it works
prompt in, output_image (IMAGE) out, POSTed to WaveSpeed's Nano Banana Pro text-to-image path. The controls are minimal and honest:
prompt- the description. Pro reasons before rendering, so a specific sentence beats a vague one.aspect_ratio- ten choices, 1:1 through 21:9, default 1:1.resolution-4k($0.15) or8k($0.18), default 4k. The whole point of this node.output_format-pngorjpeg.
enable_sync_mode defaults on.
Installing
cd ComfyUI/custom_nodes
git clone https://github.com/symbiotica-ai/comfyui-nodes.git symbiotica
pip install -r symbiotica/requirements.txt
Restart or use ComfyUI Manager (search Symbiotica), and feed an NS WaveSpeed Client node carrying your WAVESPEED_API_KEY into client.
Common issues
- 8K is a monster tensor. A multi-thousand-pixel image moving through downstream nodes will test your RAM and your patience. Only go 8K when the deliverable genuinely calls for it.
- No seed control. Same as the rest of the Pro line - reproducible output isn't offered here. The base (non-Pro) Nano Banana node has a seed if you need one.
- The model's filter is the filter. Closed Google model, unchangeable refusal behavior. No node setting bypasses it.
- Per-call metering. Iterating on a prompt with 8K re-rolls is a quiet way to spend real money. Draft at 4k, commit at 8k.
The trap to avoid is the "three cents more, why not" default to 8K. The model's imagination is the same at both sizes - 8K is for when the pixels matter, not the idea.
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| client | WAVESPEED_AI_API_CLIENT | — | |
| prompt | STRING | Text description of the image to generate | |
| aspect_ratio | COMBO | 1:1 | Aspect ratio of the generated image |
| resolution | COMBO | 4k | Output resolution: 4k ($0.15), 8k ($0.18) |
| output_format | COMBO | png | Output format - use PNG for transparency support |
| enable_sync_mode | BOOLEAN | true | Wait for generation to complete before returning |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| output_image | IMAGE | — |