🍌 普通单图生成
One prompt, one hero shot, zero GPU — the minimal Banana node
- image_1
- image_2
- image_3
- image_4
- image_5
- image_6
- image_7
- image_8
- image_9
- image_10
- image
The "🍌 普通单图生成" node is the whole Hrio-Commerce pack stripped to its smallest useful form: a prompt box, a model dropdown, an API key, and one IMAGE out the other side. If you just want a single e-commerce hero shot without the three-view or five-view machinery, this is the node you grab.
There's genuinely not much to it, and that's the point. The multi-view nodes exist because product listings need consistent angles; this one exists because sometimes you need one clean image - a hero shot, a category thumbnail, a quick comp - and you don't want a node that fires three paid API calls when you only need one.
How it works
Same skeleton as every node in this pack: an HTTP wrapper around a proprietary image-generation backend. Your prompt and up to ten optional reference images (image_1…image_10) are uploaded and sent together in a single request; the generated image comes back as an IMAGE tensor ready for Save Image or Preview. No diffusion on your machine - this is the external-API-node pattern the KB documents in external-api-nodes.md: per-call cost, your inputs leave your machine, and the model you're calling isn't one you could run locally anyway.
The default prompt is a decent starting point even if you replace the subject: "high-quality e-commerce product image, subject complete and clear, pure white background, clean lighting, realistic, single subject" (Chinese). Keep that shape - clean background, single subject, no clutter - and you'll get more of the commercial look the backend is tuned for.
Inputs
prompt- the one thing you'll actually edit.model- Banano (default), Banano Pro, or Gemini 3.5 Flash. Display names for backend endpoints; Banano is the everyday pick, Pro when you need the higher tier.api_key- required; blank reads the pack'sconfig.ini.image_size(2K default) /aspect_ratio(Auto default).
The only output is image - a single IMAGE. No info string, no mp4url, no batch. That's it.
Install
cd ComfyUI/custom_nodes
git clone https://github.com/moki-flank/Hrio-Commerce
or search Hrio-Commerce in ComfyUI Manager and restart. The pack's dependencies are just requests, numpy, and pillow (torch comes with ComfyUI) - nothing heavy, and nothing to download, because there's no local model. The real requirement is the vendor's API key, in the node field or in the [banana] section of the pack's auto-generated config.ini.
Where it bites
- The key is the whole failure mode. No API key and the node hands you a red error image as its output. Check
info-style messages in the returned image if something looks off. - It's proprietary and it phones home. This is a closed-source, credential-holding, network-calling node - the exact category the KB says to treat with suspicion since the LLMVISION malware incident. Read the code before you trust it, and accept that every prompt and reference image goes to the vendor's server.
- Cost per call. It's metered. One image per run, billed per call - fine for drafts, expensive as a brute-force iteration loop.
Need consistent front/side/back instead of a single shot? Use the three-view node - same pack, same key, same install. Want a full orbit? Five-view. This node is for when "one good image" is the entire job.
Inputs (16)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | 高质感电商产品图,主体完整清晰,纯白背景,光影干净,真实自然,单个主体。 | 单图提示词。此节点只输出一张 IMAGE。 |
| model | COMBO | Banano | 图片模型。 |
| api_key | STRING | 填入 API Key;留空时尝试读取 config.ini 的 api_key。 | |
| image_size | COMBO | 2K | 图片尺寸。 |
| aspect_ratio | COMBO | Auto | 图片宽高比。 |
| automation_payloadopt | STRING | 自动化 JSON。前端默认隐藏,只在点击节点上的自动化折叠按钮后显示;后台仍会序列化并参与运行。 | |
| image_1opt | IMAGE | 参考图 1;会上传后和提示词一起发送给图片生成接口。 | |
| image_2opt | IMAGE | 参考图 2;会上传后和提示词一起发送给图片生成接口。 | |
| image_3opt | IMAGE | 参考图 3;会上传后和提示词一起发送给图片生成接口。 | |
| image_4opt | IMAGE | 参考图 4;会上传后和提示词一起发送给图片生成接口。 | |
| image_5opt | IMAGE | 参考图 5;会上传后和提示词一起发送给图片生成接口。 | |
| image_6opt | IMAGE | 参考图 6;会上传后和提示词一起发送给图片生成接口。 | |
| image_7opt | IMAGE | 参考图 7;会上传后和提示词一起发送给图片生成接口。 | |
| image_8opt | IMAGE | 参考图 8;会上传后和提示词一起发送给图片生成接口。 | |
| image_9opt | IMAGE | 参考图 9;会上传后和提示词一起发送给图片生成接口。 | |
| image_10opt | IMAGE | 参考图 10;会上传后和提示词一起发送给图片生成接口。 |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| image | IMAGE | — |