FocusBerry: class focus, without cameras

FocusBerry 🍓

Class focus that ripens in real time. No cameras, no recordings, no student data.

No cameras · No recordings · No transcription

Live focus score

Blends recent anonymous check-ins (70%) with how well the room’s sound fits the current activity (30%). The berry ripens as the class locks in.

–

Waiting for a session

Start a session and tap a few check-ins.

Session

No session running

Lesson phase & topic

Tap when the activity changes. Add a topic note first if you want it in the data. Topics come from you, never from listening to students.

Focus pulse

Anonymous 1–5 check-ins: quick hand poll, or students tap on the way past.

–

avg, last 5 min

check-ins

0:00

elapsed

Room sound (optional)

Analyzed on this device only: a loudness number plus a rough texture estimate from acoustic patterns, never words.

Off

The microphone signal is analyzed in your browser and discarded instantly. No audio is recorded, stored, or sent anywhere, and speech is never converted to text. The app keeps one loudness number and one texture label (quiet / one voice / many voices) every 15 seconds.

This session so far

Dashboard

Where focus ripens and wilts: by minute, phase, and day.

Ripeness strip

One cell per minute of class, colored from unripe to ripe by average focus.

1: wilted 2 3 4 5: ripe 🍓 no check-ins

Most recent day in range

Focus (berry red, 1–5) and room sound (seed gold, 0–100) through the class period.

Focus by lesson phase

Focus by day

What the data says

Privacy by design

No cameras, ever. Attention scoring from video is biometric surveillance of students. This tool measures the room, not faces.
No recordings and no transcription. The optional sound analysis computes a loudness number and a coarse texture label (quiet / one voice / many voices) from acoustic energy patterns, in the browser, and discards the audio instantly. Speech is never converted to text; what anyone says is never analyzed.
Topics come from the teacher. Discussion topics in the data are one-tap notes the teacher types, not anything extracted from students’ conversations.
Anonymous by default. Focus check-ins are a number and a timestamp. No names, no accounts, no way to trace a vote to a student.
Data stays on the teacher’s device. Everything lives in this browser’s local storage. There is no server. Export to CSV or clear it any time.
Humans judge, AI assists. The class interprets its own patterns. No algorithm labels any student.

Run it on GitHub Pages

Create a repository (e.g. focusberry) and upload index.html.
Settings → Pages → Deploy from a branch → main, folder / (root).
Your site appears at https://<username>.github.io/focusberry/ in about a minute.

Tip for presenting: open the Dashboard and press “Load sample week” so the charts are full, then do the live demo. The berry ripening on screen while the room votes is the moment.