Observed arrival · 2026-10-23
PatchCore teaches cameras and microphones what “normal” looks like
An open-source visual and acoustic anomaly detector that learns from normal examples and runs locally in a browser or native software.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- People monitoring repeatable visual or acoustic processes
- Worth noticing
- Its own guidance says scores measure unfamiliarity, not fault probability or cause; consistent framing and microphone placement help.
Field notes
The workflow separates collecting normal examples from calibrating a threshold with additional normal observations, then presents live scores alongside visual heatmaps or sound spectra. The page lists version 0.2.0 downloads and model files with license and provenance links; native capture can use FFmpeg, with optional Aravis and GStreamer integrations. Browser capture requires HTTPS and permission, and the tab must remain visible during inspection.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue