rubenppombo/birds-app
🦅 IberBirds
IberBirds is a tool to help identify birds of prey and storks commonly found in the Iberian Peninsula. I built this project to explore how computer vision can assist in nature observation without complex infrastructure.

About the Project
The concept is simple: you upload a photo of a bird, and the application identifies it along with a confidence score. It runs completely in the browser/cloud session—no accounts, no databases, and no tracking. Once the analysis is done, the image is discarded.
I trained a YOLO11 model specifically for this task, focusing on 10 distinct species often confused by casual observers (like differentiating a Red Kite from a Black Kite).
How it runs
Unlike typical web apps that separate frontend and backend on different providers, this project is packaged as a single Docker container.
- Frontend: A React application (bundled with Vite) handles the user interface.
- Backend: A lightweight FastAPI Python server hosts the model and serves the static frontend files.
- Inference: When you upload an image, it's processed in-memory by the YOLO model running on the server.
You can try it live here: [rubenppombo-birds-app.hf.space](https://rubenppombo-birds-app.hf.space)
Supported Species
The model detects the following birds:
- Storks: White Stork, Black Stork.
- Vultures: Griffon Vulture, Cinereous Vulture, Egyptian Vulture.
- Raptors: Spanish Imperial Eagle, Golden Eagle, Red Kite, Black Kite, Peregrine Falcon.
Running Locally
If you want to run this on your own machine (requires Docker):
# Clone the repository
git clone https://github.com/Rubenpombo/birds_app.git
cd birds_app
# Build and run
docker build -t iberbirds .
docker run -p 7860:7860 iberbirdsThe app will be available at http://localhost:7860.
