Inspiration'
We are a team of nature lovers and explorers. Every time we would go on an outdoor adventure, I noticed we spent more time trying to take photos and keep track of the animals we were seeing, instead of actually enjoying the moment. Then, we would go home and attempt to research the animal, which proved to be impossible because we had already forgotten all the details. We wanted a way to get the information without having to interrupt the experience.
What it does
TailMate is an intelligent field assistant that automates the documentation of reptile sightings. Using a combination of ultrasonic sensors and Vision AI, it detects when a reptile is in range, snaps a high-resolution screenshot from a camera feed, and identifies the species. It then automatically syncs local weather data (temp/humidity) and GPS coordinates to generate a "Field Log" entry, providing instant conservation status and safety advice.
How we built it
Hardware: Arduino-based sensors (ultrasonic, DHT11) detect proximity and environmental data.
Intelligence: Integrated the Gemini 3 Flash API to handle complex species identification and habitat analysis.
Backend: A Flask server bridge coordinates the local hardware, the AI, and the cloud.
Database & Storage: Supabase handles our real-time data logging and image hosting.
Frontend: A responsive, "paper-style" journal UI built with HTML5, CSS3, and JavaScript, featuring Leaflet.js for geographic mapping.
Reference Database: We utilized the iNaturalist API to pull a baseline of reptile observations specifically for San Diego County. This gave TailMate a "knowledge base" of local species, their typical habitats, and conservation statuses.
Challenges we ran into
The biggest hurdle was the "Camera Conflict." We had to design a logic system that could maintain a smooth, live video stream for the dashboard while simultaneously "hijacking" the camera for a split second to take a high-res snapshot for the AI without crashing the feed or losing the frame. Another challenge we ran into was with the data. Because our computer capabilities were limited, we had to reduce the dataset to include only reptiles found on campus.
Accomplishments that we're proud of
We are proud of the automated logging system we have integrated into the application. The ability to update the log with the current statistics of where you are at required us to put in a lot of front-end and backend work.
What we learned
We learned the intricacies of Edge-to-Cloud architecture. Managing the communication between physical sensors and cloud-based AI taught us how to optimize data packets and handle asynchronous API calls to keep the user interface feeling fast and responsive.
What's next for Tailmates: your field note helper!
Offline Mode: Implementing local "lite" models for deep-woods exploration with no cell service. Motion sensors: detecting and
Log in or sign up for Devpost to join the conversation.