01
Start here
A listening camera
Let the place
make the picture.
Your browser microphone uses local YAMNet for everyday sounds. ESP32 recordings use the Perch 2.0 backend for detailed animal species.
Preparing the on-device classifier…
02
Loading sound model
Record
00:20
LIVE AUDIO0%
DETECTION RESULTS
Ready after locationSound events appear here while you record.
Waiting for sound…
ESP32 · PERCH 2.0 BACKEND
Idle
No live ESP32 job yet.
Browser YAMNet stays local · ESP32 audio is sent to the Perch backend
03
Waiting for recording
FLUX souvenir
✦Your illustrated place appears herelocation + sound → prompt → FLUX image
FLUX is painting your soundscape…
✦Your scene will emerge heresound → labels → image
A listening placeFLUX souvenir
HEARD HERE
IMAGE PROMPTlocation + landmark + detected sounds
1
The microphone is split into one-second windows. No audio file is saved.
2
Browser audio uses local YAMNet. ESP32 audio uses backend Perch 2.0 for detailed animal names.
3
Location, a real landmark and visible sound sources generate a landscape souvenir.