AI bird ID rendered as hand-cut 1800s natural-history plates

What it does It sits between a microphone and a display, polling BirdNET-Go for real-time species detections and matching each identified bird to a hand-curated, background-removed illustration from 19th-century natural-history plates. The result is rendered either on a Pimoroni Inky Impression e-ink panel or as a web kiosk, packing birds onto a textured page sized roughly by body mass.

The interesting bit This is essentially a very opinionated, aesthetic display layer for BirdNET-Go. The author has manually cut out over 800 illustrations covering more than 400 species—no AI-generated art here—and the layout algorithm places larger species toward the center of the page. When no birds are present, the frame shows a bare perch, which is a nice touch.

Key highlights

• Fully local AI detection via BirdNET-Go; no cloud required.
• Over 800 hand-curated, public-domain illustrations from real 1800s plates.
• Supports both e-ink (Pimoroni Inky Impression) and standard web kiosk displays.
• Can point to an existing BirdNET-Go instance on the network, or run alongside it.
• Admin page included for configuring what appears on the frame.

Caveats

• Still in early development; the README warns of odd bugs and unpolished edges.
• Illustration coverage is strongest for Scandinavia, the British Isles, and Germany; other regions are sparse.
• Requires a separate BirdNET-Go installation (or container) to handle the actual audio classification.

Verdict Birdwatchers with a Raspberry Pi and a taste for natural-history aesthetics will love this. If you are looking for a generic dashboard or do not care about 19th-century illustration, it is probably overkill.