Why it’s interesting
Fugleramme turns birdsong outside a window into a changing natural-history collage. A microphone feeds the locally running BirdNET-Go classifier, then the project selects matching birds from more than eight hundred hand-cut historical illustrations. Birds are scaled using real body mass and dithered into the six colours of an e-ink panel. The creator is explicit that no generative image model is involved: AI identifies calls, while the pictures come from traceable nineteenth-century plates. The frame refreshes only when the detected set changes, making it feel more like a calm window than a dashboard.
How to try it
The simplest route is the live web view from the creator's garden in Bergen, Norway; it needs no account and shows current detections with their collage. To reproduce it, run BirdNET-Go with a microphone and connect the provided container. The creator uses a Raspberry Pi 5 and a 13.3-inch Inky display, but the panel is optional and web-only mode works. Artwork coverage is currently strongest for the Nordics, Britain, and Germany, with thinner North American coverage.
Requirements and sources
Arne Giacomo Munthe-Kaas publicly launched the project on September 15, 2026. Code is MIT-licensed, the classic artwork directory is CC BY-SA 4.0 with a source manifest, and BirdNET-Go has its own CC BY-NC-SA 4.0 terms. Building the physical frame requires bought and assembled hardware, so this entry is marked paid and device-required; the live view remains free. The repository's official project image is used only for identification.
[Project and build guide](https://github.com/arnegiacomo/fugleramme) · [Live bird frame](https://fugleramme.arnegiacomo.dev/)