Picsart Image Vectorizer
Turn raster art into a real SVG
- image
- file_path
The Picsart Image Vectorizer is the pack's quiet utility node: you feed it a raster image, it hands you back a vector file on disk. No prompt, no seed, no diffusion. It's a conversion service, Picsart's model running on Runware's cloud, and it sits in the pack because "every Runware model as a node" includes the boring ones that are still useful.
Why bother? Vector output means infinitely scalable art - a logo you generated at 1024px can become a sign the size of a building without pixelating. It's the classic endgame for AI-generated logos, stickers, and illustrations: the model draws it, the vectorizer makes it production-ready. That's the workflow this node is built for, and it's a one-node affair.
What you actually set
Almost nothing. The only required input is image - an IMAGE socket, so any image you've generated or loaded in ComfyUI works. In the optional list:
outputFormat-SVG. That's the entire list of choices, because vectorizing to anything else would defeat the point.ttl- if you ever need the cloud-side URL to live longer than the default 60 seconds. Ignore it in 99% of cases.
There's exactly one output: file_path, a STRING pointing at the SVG the node saved into your ComfyUI output folder. That's the interesting quirk - unlike image or video nodes that return native types, vector output comes back as a saved file path, because ComfyUI has no native SVG tensor. You can't preview it in the node graph directly; open the file path, or wire it into a text display to see where it landed.
Install
Same one-time pack setup: ComfyUI Manager → search Runware → install → restart, or
cd ComfyUI/custom_nodes
git clone https://github.com/Runware/ComfyUI-Runware
pip install -r ComfyUI-Runware/requirements.txt
API key from runware.ai/api-keys into ComfyUI Settings → Runware API key or RUNWARE_API_KEY. The pack's only dependencies are runware-sdk, pillow, and soundfile; no models, no downloads.
Where people get burned
- Vectorization is not magic. A complex photographic image turns into an SVG that's huge, layered, or just ugly. Feed it clean flat-color art - logos, icons, simple illustrations - and it shines. Feed it a photo and you'll be disappointed.
- The output is a file path, not an image. People wire it into an image preview out of habit and get nothing. Read the
file_pathoutput instead. - It still costs a run. It's a cheap run compared to generation, but the title bar cost readout won't be zero.
Honest verdict: this is a niche node you install once, need occasionally, and are glad exists on the day a client asks for a logo in SVG. Not a daily driver - a tool in the drawer.
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | — | |
| 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 | SVG | File format for the vectorized output. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| file_path | STRING | — |