Preview FeatureEffect ⚡🅡🅞🅣🅘

The whole point of this pack's FEATURE system is that a lot of nodes get driven by a numeric signal you can't directly see - audio amplitude, brightness, depth variance, a hand-drawn curve. That's powerful, but it's also a black box: is the signal actually doing what you think it's doing at frame 47? This node is the debug tool for that exact question. It burns the raw feature value as text directly onto your video frames, so you scrub through the output and watch the number next to the footage it's describing.

How it works

Feed it your video_frames and the feature you want to inspect, and for each frame it overlays the feature's current numeric value as text, using whatever position, size, and color you set. It's not part of your actual effect chain - you drop it in as a side-branch to sanity-check a feature before (or while) you use it to drive something else, then remove it once you're confident the signal is doing what you expect.

The inputs and outputs that matter

• video_frames (IMAGE, required) - the frames to overlay the label onto.
• feature (FEATURE, required) - the signal you want to inspect.
• text_color (STRING, default "(255,255,255)") - the label's color.
• font_scale (default 1, range 0.1–2) - label text size.
• label_prefix (STRING, default empty) - optional text prefixed before the number, useful when you're comparing multiple visualizers stacked in one workflow so you can tell which feature is which.
• label_x_offset / label_y_offset (default 10, range 0–1000) - where the label sits on the frame.

Output: a single IMAGE - your video frames with the feature value burned into each one as text.

How to install it

Via ComfyUI Manager: search "RyanOnTheInside," install, restart. By hand:

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

then restart. Pure overlay/rendering work on frames you already have - no model download.

Common issues & troubleshooting

Text is unreadable against the footage. Adjust text_color to contrast with your scene, and bump font_scale if it's too small to read at your output resolution. There's no automatic outline or background behind the text, so busy footage can genuinely swallow a light-colored label.

Comparing two features and can't tell which label is which. Use label_prefix on each EffectVisualizer instance - something like "brightness: " vs "beat: " - and offset their label_x_offset/label_y_offset so the two overlays don't sit on top of each other.

Values on screen don't match what you expected from the source. That's a real finding, not a bug in this node - it's telling you honestly what the feature is outputting. Trace back to whichever Feature-source node produced it (frame rate mismatch, wrong extraction method, unintended clamping) rather than assuming the visualizer is wrong.

Forgot this is still in the graph after finishing debugging. Since its output is a modified copy of your video with text burned in, leaving it wired into your final output by accident will show up as an unwanted number overlay in your actual render - disconnect or bypass it once you're done inspecting.