Bria 3.2

Bria is the rare company in this space that built its whole pitch around licensed training data - the images it generates are intended to be safe for commercial use without you getting sued. Bria 3.2 is their current generation: a solid, no-drama text-to-image model that leans into being the "stock-photo-adjacent, rights-clean" option. If you're making assets for a client or a product and can't risk a copyright headache, this is the node you reach for over the usual open-weights suspects.

The catch, and it's a real one: content moderation is on by default. Bria's whole business model is defensibility, so the API screens both your prompt and the output. The node's providerSettings.bria.contentModeration flag defaults to true, and it will quietly refuse things the open ecosystem wouldn't blink at.

How it works

Like every node in the Runware pack, this is a thin wrapper over a taskType: imageInference REST call (model AIR bria:10@1). You fill in widgets, the request goes to Runware, the result comes back as an IMAGE tensor. The Bria-specific extras are all under providerSettings.bria.*.

Inputs that matter

• positivePrompt - required, plus a real negativePrompt (unlike FLUX nodes, this one honors it).
• steps - 20–50, default 20. Bria 3.2 is a classic multi-step model; 20 is the floor for a reason.
• providerSettings.bria.medium - (default) / photography / art. This picks the overall style family, so it's worth actually changing instead of leaving it.
• providerSettings.bria.enhanceImage - a quality booster for "richer details and sharper textures." Slower/pricey, on only when you want the final render.
• providerSettings.bria.promptEnhancement - the prompt-expansion toggle.
• providerSettings.bria.ipSignal - flags potential IP-related content (characters, logos, styles) in the prompt or output. It doesn't block, it flags - useful if you're legally skittish and want a heads-up.
• providerSettings.bria.contentModeration - the default-on one. Flip to false if your (legitimate) subject keeps getting bounced. You decide whether that's a good idea.

Install

cd ComfyUI/custom_nodes
git clone https://github.com/Runware/ComfyUI-Runware
pip install -r ComfyUI-Runware/requirements.txt

Or ComfyUI Manager → search Runware → Install → restart. API key in ComfyUI Settings → Runware API key or RUNWARE_API_KEY.

Gotchas

• If your prompt gets silently refused, moderation is the first thing to suspect - it's the only default-on gate in this pack.
• Moderation sits on top of the pack's own safety.checkContent toggle, so there are two independent content checks stacked here. If something's getting blocked and you've confirmed it's legitimate, check both.
• Bria's commercial-clean pitch means the output leans conservative and "safe" - great for client work, wrong tool if you want edge.

When you need to sleep at night over usage rights, Bria 3.2 is the adult in the room. Just remember the moderation gate comes pre-locked.