Meux Artistic Text Preview

Text on images is one of those things you usually leave ComfyUI for - render, open Photoshop or GIMP, add the title, come back. If you're producing thumbnails, logos, or title cards in a batch workflow, this node exists to kill that round trip: it renders styled text on a transparent canvas and hands you both the image and a matching mask, so you can composite it straight onto your render.

It ships in the Baidu Meux ComfyTools pack, and it's easily the most feature-dense node in it. Fill, outline, shadow, inner shadow, glow, gradients, per-character spacing, alignment - the kind of surface area you'd expect from a dedicated text tool, not a utility node.

How it works

The node renders your text to a text mask with Pillow, then builds up layers - fill, outline, effects - and composites them over an optional background. It outputs two things: an RGBA IMAGE preview and a MASK derived from the text's alpha. Keep background_opacity at 0 and the image comes out fully transparent, ready for compositing.

Fonts are the interesting bit. The node looks in the pack's own fonts/ folder first, so drop in any font file there and it appears in the font_name dropdown. If it's not there, it falls back to bundled and system fonts. The bundled defaults are Alibaba PuHuiTi variants, a CJK typeface family - perfectly usable for Latin text, but if you want a specific look, add your own font and refresh.

Two font sizing modes: fixed uses the literal font_size; fit scales the text down to fit the canvas. It won't scale up past your font_size, so think of that as the cap.

The inputs that matter

• text - multi-line supported, newlines and all.
• font_name - dropdown, populated from your fonts/ folder first.
• width / height - the output canvas, up to 4096.
• font_size_mode / font_size - fixed vs fit.
• fill_type - solid, gradient, or none; for gradients pick fill_color_1, fill_color_2, and a fill_direction.
• outline_type / outline_width - solid or gradient stroke around the glyphs.
• effect_preset - one-click soft_shadow, heavy_shadow, soft_glow, neon_glow, engraved, or shadow_glow, with effect_strength as a multiplier.
• background_opacity - 0 means fully transparent.

Outputs: image (the rendered IMAGE) and mask (the text alpha). Wire image into a composite node over your render, or save both.

Installing it

Pack install, same as everything in this set:

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

Or use ComfyUI Manager (search "Baidu Meux ComfyTools"). Restart, then add your fonts:

cp MyFont.ttf ComfyUI/custom_nodes/Comfyui_MultiSaveImage/fonts/

Restart (or reload) and the font shows up in the dropdown.

Where people get burned

• Font doesn't appear. It's cached, or it's in the wrong folder - the pack's fonts/ directory, not ComfyUI's.
• Defaults feel Chinese-first. The bundled defaults are PuHuiTi weights. Nothing wrong with them for Latin text, but that "why is my default font so heavy" reaction is usually just the Alibaba Heavy/Black default doing its thing.
• Mask looks hard-edged when composited. The mask comes from the text alpha, so it's sharp by design. If you want a soft edge, blur it downstream.

If you do text-on-image work, this node is the reason to install a pack you might otherwise skip. It's the rare ComfyUI text node that doesn't feel like a toy.