Kandinsky 5.0 Image Lite
The Runware node that actually feels like local ComfyUI
- lora
- acceleratorOptions
- advancedFeatures.watermark.image
- image
Most Runware nodes are deliberately dumbed-down: prompt in, image out, two widgets. Kandinsky 5.0 Image Lite is the exception that proves the rule - it's the one cloud node that hands you the full diffusion control panel. Steps, scheduler, CFG, negative prompt, even a LoRA socket. If you're coming from local ComfyUI and want the same feel without the GPU, this is the bridge.
It's the hosted "Lite" tier of the Kandinsky image model family (Kandinsky is the long-running diffusion line from the Russian Sber-backed labs), served through Runware. The Lite suffix means the fast, economical variant - which, given the control surface, is the sweet spot for people who miss tuning a sampler.
How it works
Required: positivePrompt, width, and height (both default 1024, stepped in multiples of 16 - keep them divisible by 16 like you would locally).
Then the controls that make this node feel like home:
steps- 10 to 50, default 20. More steps, more detail, longer wait and higher bill.scheduler- 13 choices including the DPM++ Karras family, Euler / Euler a, LCM, DDIM, DDPM, and PNDM. This dropdown alone is why you'd pick this node over a plain txt2img.CFGScale- 2 to 10, default 7. How hard the model clings to your prompt. 7 is a fine starting point; 10 will overbake, 4–5 gets freer.negativePrompt- with CFG above 1 it's respected; the tooltip notes it's ignored when guidance is disabled.seed- full 64-bit reproducibility when you lock it.lora- aRUNWARE_LORAsocket fed by the Runware LoRA builder, so you can stack a community LoRA onto a hosted run. That's a genuinely unusual move for a cloud node and worth knowing about.acceleratorOptions- aRUNWARE_ACCELERATOROPTIONSsocket (from the Runware Accelerator Options builder) if you want to play with caching tricks.
Output is a native IMAGE, so it drops into Save Image or feeds an upscale pass like any local node.
Installing it
Standard official pack install:
cd ComfyUI/custom_nodes
git clone https://github.com/Runware/ComfyUI-Runware
pip install -r ComfyUI-Runware/requirements.txt
Restart, set your API key in ComfyUI Settings → Runware API key or via RUNWARE_API_KEY. Dependencies: runware-sdk, pillow, soundfile. No model downloads - the checkpoint is Runware's.
Where people get burned
Two traps. First: you get more controls here, so you'll fiddle, and every extra step is metered - a 50-step run with CFG 10 costs noticeably more than the defaults, and this is a paid top-up API with the pack's standing $20-minimum complaint. Second: the "Lite" naming - it's the economical tier, not the flagship; don't expect it to out-render a dedicated high-end model. And the usual pack rule: no API key set means instant errors, so check Settings before you blame the scheduler.
Inputs (24)
| Name | Type | Default | Description |
|---|---|---|---|
| positivePrompt | STRING | Text prompt describing elements to include in the generated output. | |
| loraopt | RUNWARE_LORA | — | |
| acceleratorOptionsopt | RUNWARE_ACCELERATOROPTIONS | — | |
| advancedFeatures.watermark.imageopt | IMAGE | — | |
| negativePromptopt | STRING | Prompt to guide what to exclude from generation. Ignored when guidance is disabled (CFGScale ≤ 1). | |
| sizeopt | COMBO | 1K (1:1) | Output size. Only one of resolution or dimensions is sent. |
| seedopt | INT | 00–9223372036854776000 | Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range. |
| stepsopt | INT | 2010–50 | Total number of denoising steps. Higher values generally produce more detailed results but take longer. |
| scheduleropt | COMBO | (default) | Scheduler to use for the diffusion process. |
| CFGScaleopt | FLOAT | 7.002–10 | Guidance scale representing how closely the output will resemble the prompt. Higher values produce results more aligned with the prompt. |
| numberResultsopt | INT | 11–20 | Number of results to generate. Each result uses a different seed, producing variations of the same parameters. |
| advancedFeaturesopt | BOOLEAN | false | Enable to set advancedFeatures. Off uses the model's default. |
| advancedFeatures.watermark.bgColoropt | STRING | Background color in hex format. | |
| advancedFeatures.watermark.displayPositionopt | COMBO | (default) | Watermark position. |
| advancedFeatures.watermark.fontColoropt | STRING | Text color in hex format. | |
| advancedFeatures.watermark.opacityopt | BOOLEAN | false | Enable to set advancedFeatures.watermark.opacity. Off uses the model's default. |
| advancedFeatures.watermark.opacity_valueopt | FLOAT | 0.100.1–1 | Watermark opacity. |
| advancedFeatures.watermark.textopt | STRING | Watermark text. | |
| safetyopt | BOOLEAN | false | Enable to set safety. Off uses the model's default. |
| safety.checkContentopt | BOOLEAN | false | Enable or disable content safety checking. |
| ttlopt | BOOLEAN | false | Enable to set ttl. Off uses the model's default. |
| ttl_valueopt | INT | 60 | Time-to-live (TTL) in seconds for generated content. Only applies when `outputType` is `URL`. |
| outputFormatopt | COMBO | JPG | File format for the generated image. |
| outputQualityopt | INT | 9520–99 | Compression quality of the output. Higher values preserve quality but increase file size. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| image | IMAGE | — |