keenie

Craft better experiences

Predict retention before you publish your video. Keenie uses attention estimation and visual engagement metrics to show when your audience loses focus—using a standard webcam, no special hardware.

Powered by MediaPipe AI
Privacy First - On-Device Processing
Standard Webcam - No Special Hardware

The Problem with Traditional Analytics

YouTube retention graphs only tell you what happened, not why. keenie shows you the why.

Stop Guessing

No more wondering why viewers drop off. See when attention drops and when they re-engage—aligned with your content timeline.

0:005:0010:00

Attention Estimation

Estimate relative gaze direction and screen quadrant focus. Detect on-screen vs. away, look-away duration, and time spent engaged—using a standard webcam.

Facial Expression Analysis

Non-psychological: smile intensity, eye openness, blink rate, and expression variance over time. Detect attention-drop and re-engagement moments.

ExpressiveHigh intensity
NeutralLow variance
Attention dropRe-engagement moment

How It Works

Get actionable insights in three simple steps

01

Upload or link your draft

Drop your video file directly into keenie or simply paste a private link from YouTube or Vimeo to get started.

02

Privacy-first testing

Send a secure link to your audience. Testers grant camera access for local-only AI processing. Gaze and facial expression data are calculated on their device—no raw video is ever recorded or uploaded.

03

Watch the insights flow

Analyze engagement scores (0–100), frame-by-frame attention heatmaps, and emotional reaction timelines. Compare different versions side-by-side to see which hook wins.

Frequently Asked Questions

Everything you need to know about keenie’s predictive analytics.