Open Age Detection

This is the smallest node in the Runware pack and possibly the most honest one. Runware_open_age_detection takes an image, looks at the face(s) in it, and returns an age estimate as text. That's it. One required input (image), no optional knobs at all, and a text output (a STRING). It's a vision utility (runware:154@1, task type caption) doing age estimation in the cloud, which is a genuinely niche capability to have available as a single ComfyUI node.

When would you actually use it? It's a quality-control and validation tool, not a creative one. If you generate portraits at scale - say, casting or character-consistency work where the model is supposed to render "mid-30s" - you can route a batch through this node and check whether the faces are landing in the expected age band, instead of eyeballing every image. It's the same role a tagger or an NSFW-check plays in a batch pipeline: an automated gate that catches outliers before they reach a human.

How to use it

The workflow is almost absurdly simple: wire any image output into image, queue, and read the age estimate from the text output. That string can feed a Display Text node if you want to see it on the canvas, or go into any downstream logic that consumes strings. Since the pack's text nodes return plain STRING, you can chain the result into a conditional or a batch-report node without any type gymnastics.

Because it runs on Runware's cloud, there's no local age-detection model to install, no OpenCV pipeline to maintain - just the pack, which you already have if you're reading this page. Install is the pack's standard: ComfyUI Manager → Runware, or git clone https://github.com/Runware/ComfyUI-Runware into custom_nodes/, plus an API key in ComfyUI Settings.

The honest take

• It's a single number from a black box. Age estimation from a face is approximate, biased, and sensitive to image quality - a blurry or stylized portrait will give you noise. Treat the output as a rough check, not a measurement.
• No tuning knobs means no surprises. You can't adjust confidence or region; you get the model's best guess. That's fine for a utility, and it keeps the node trivial to use.
• It's cheap. A per-run cost appears on the title bar, and for a caption/estimate task it's about as small as this pack gets.
• Real-world fit - this is the node you use in a batch automation, not the one you show off. But batch automation is a legitimate reason to run ComfyUI at all.

If you've ever wanted to automatically gate generations by apparent age or just had the idle question "how old does the model think this face is," this is a $0.00-class experiment that's already wired into your graph.