Home
Snapshot's animal check and score threat
Animal Check
Camera options
Feed a frame
Dashboard
Live Camera options
Starting Page
Ring
Live Camera
Ring MCP

Inspiration

I came across a Reel that talked about Ring — the company, before it was sold to Amazon. Watching it led me to learn more about what Amazon Ring does. This was actually my first time learning what Ring is and some of its features. While searching Google for Amazon devices and reading people's reviews, I came across a tweet about an Amazon device. When I opened it, I discovered it was actually a hackathon for Ring. That's when I decided I wanted to understand the Ring ecosystem and build something meaningful for it.

What it does

It helps the user, integrate the camera, click a live snapshot and get the result which consists of danger score, and animal name. We have integrated all user cameras, one place, based on different farm location. Live feed — watch any camera live (WHEP) via the Ring Partner API. Upload & identify — upload any picture or frame and it identifies the animal and how critical it is, blended with the farm's region, camera zone, and time of day (a raccoon at midnight near the coop ≠ a neighbor's cat at noon) We also provide detection history, emergency contact information and help user consolidate the information.

How we built it

We started as a mono Repo First making CI/CD standard pipeline so that it becomes easy to deploy it on docker, and EC2. Later i integrated Vercel and Deployed it there. We have used TypeScript, and JavaScript to build our application layer, used Drizzle, Open Code and ChatGPT to learn and then integrated our database S3 storage and Open MAPS API for Nearest Location Police and Emergency situation. Using ChatGPT documentation i have also made it easy for my website to be crawled using AI agents, created files, relevant to Agentic- ease of Crawl.

Challenges we ran into

In this project, I learned some things the hard way: deployment. First, I used Docker and EC2. Since I didn't have a public URL and wanted to keep costs low, I moved my deployment to Vercel. This was new to new and i spend a lot of my initial time figuring this our. I then switched to neonTech which offers us S3 and RDS type storage and DB but at ease to first time users and no charges at certain costs. Once my Production code failed but my local pre-production was successful working, The challenge was i missed my Database upgrade properly. For me, TypeScript was still new; I mostly had hands-on experience with Java and C++. But I was learning it, and building this project helped me a lot — it improved my logical reasoning along the way. I also understood, together with my teammate, how to use S3 and move URLs into the database to make everything less overwhelming.

Accomplishments that we're proud of

Confidence that i can move, and even maintain a code quality, with real production checks using proper CI/CD builds. Taking from Ideation to Production is a huge booster for me. Not that i have not taken earlier but this time i did it with maintained Pipeline, keeping in mind that many people can work on it so created stable DB, pipeline Credentials.

What we learned

Understanding and thinking from consumer prospective. i checked Reddit and google first to understand if there is any issues and complaints against the product that i can understand and build Learned a lot about MCP initially my idea was to expose this using our MCP. so i also got a hands on tools, context and Agentic friendly websites.

What's next for Predator Guard

Proper Live Video Feed Integration. Integration of an good opensoruce model for proper identification Automatic Alerts on Phone. Weekly update and Suggestion to keep farm home safe during a time of year. Integration of Fire TV for live monitoring