🧪 SAMPLE REPORT · SYNTHETIC DATA NOT A REAL PERSON OR ACCOUNT. Every figure below is invented for illustration; your real report is generated locally from your own sessions. ← swear-jar home
🫙swear-jar / the damage report
GitHub processed locally
A developer tool for Claude Code + Codex

Swear Jar scans the AI prompts and recordings you already have on your machine, counts the swear words, and turns the results into a playful engineering report. It never uploads transcripts. The report is generated locally; only an explicit share or leaderboard opt-in can send disclosed aggregate swear statistics.

GOLD STAR — more please-and-thank-yous than swears. Who ARE you?
rage.wavREC · 00:00 / ∞
side a · rage.wav
$0
owed to the jar.

The usual suspects

Words caught on the mic

most-used swears · by severity
Corruption evidence

You vs. the machine

who's been swearing — you, or the assistant you corrupted
You
The machine

The whole point, obviously

Robot uprising survival odds

every coin lowers them; clean days claw them back
Scene of the crime

Where the swearing happens

coins by project · top 10
The tape, over time

Rage timeline + rage-o-clock

when you swear, and how hard it lands
Daily volume
Hourly volume + normalized frequency

The weekly pattern

By day of week

which day the machine dreads
Notable takes

For the record

Also on the tape

The fine print

what else your mouth gave away
Release the tape

Only the number leaves — never your words

aggregate figures only · your transcripts stay here
🫙 rage, wrappedrage.wav · 2026
0
processed locally · nothing left this machine · swearjar.unfocused.ai
🫙 Feel guilty? Donate to the jar