Inspiration
Running a software development agency taught me that most AI automations are passive — they only act when a user asks them to. I wanted to build something proactive.
Then two things hit me: I visited a shop using AI-integrated CCTV to monitor customer behavior in real time — and I realized almost no one was bringing that power to small businesses. Then scrolling online, I noticed barely any AI tools existed for Arduino builders — even though I had already built an AI-powered autonomous car and drone myself.
CEOVA was born from all three: agency experience, real-world observation, and a builder's frustration.
What it does
CEOVA is an AI agent platform that actively runs your business — without waiting to be asked.
Currently working: → Connects Gmail, Google Calendar & Drive with an AI agent that reads, decides and acts → Automates tasks and makes decisions on behalf of the user → Runs locally via Ollama so users can reduce API costs to near zero
Coming very soon: → CCTV camera monitoring with behavior analysis → Billing machine integration → AI code generation for Arduino Uno/Nano builders → Full offline-to-online business integration
How we built it
Built solo at 19, using TypeScript, Node.js, Google APIs, and LLMs via both cloud and local Ollama models.
The architecture separates cloud integrations (Gmail, Drive, Calendar) from the offline layer (CCTV, hardware, billing) through a unified AI agent brain that plans, decides and executes tasks autonomously — not reactively.
409+ commits in April 2026 alone across 5 repos.
Challenges we ran into
The hardest problem was bridging offline devices like CCTV cameras and billing machines with an online AI system in real time.
Unlike cloud APIs that have SDKs and docs, offline hardware integration means handling inconsistent data streams, latency, and zero standardization — all while keeping the AI agent's decisions accurate and fast.
Accomplishments that we're proud of
→ Built a fully working AI agent connected to Gmail, Calendar and Drive — solo → Enabled local AI execution via Ollama, making CEOVA accessible to users with no API budget → Designed an architecture that scales from cloud tools to physical hardware like CCTV and Arduino — no one else is doing this → 30+ GitHub repositories shipped at 19 years old while running a dev agency
What we learned
That the biggest gap in AI isn't intelligence — it's integration. Most businesses, especially in India, live half-online and half-offline. AI tools ignore that reality completely.
We also learned that local models like Ollama are underrated — giving small businesses real AI power without cloud costs.
What's next for CEOVA — Your Business's AI Brain
→ CCTV integration with customer behavior AI → Billing machine + POS system connectivity → AI code generator for Arduino Uno/Nano builders → Expanding to physical sensors and IoT devices → Launching to real businesses across India → Making CEOVA the operating system for any business — online, offline, or both
Built With
- ai-agents
- arduino
- css
- github
- google-calendar-api
- google-drive-api
- google-gmail-api
- html
- javascript
- llms
- next.js
- node.js
- ollama
- openai-api
- postgresql
- python
- react
- rest-apis
- supabase
- typescript

Log in or sign up for Devpost to join the conversation.