NEOVRY13 Agentic Engine

An autonomous AI agent powered by Gemini and Node.js that automates script analysis, lead scoring, and background CRM workflows in real time.

Comment

Inspiration

High infrastructure costs and complex server setups prevent small business owners, affiliate marketers, and solopreneurs from running 24/7 AI-driven sales pipelines. We wanted to build a zero-server-cost solution that runs autonomous AI agents directly on mobile and edge devices.

What it does

NEOVRY13 OS is an agentic CRM operating directly on edge hardware. It automatically evaluates customer intent, scores inbound leads, manages data persistence, and triggers personalized multi-channel follow-ups using a hybrid architecture of cloud-based LLMs and local models.

How we built it

  • Runtime & Process Management: Deployed on Termux with PM2 to maintain persistent background execution.
  • AI Orchestration: Integrated Google Gemini API for complex lead qualification alongside local LLMs (Ollama) for offline tasks.
  • Database & Persistence: Utilized SQLite and Supabase for local logging and real-time backend synchronization.
  • Automation Engines: Built custom Python modules (agent.py, sentinel_neovry.py) for real-time CRM monitoring and tool execution.

Challenges we ran into

Running complex multi-agent workflows and maintaining persistence on mobile Android environments without dedicated cloud servers required careful process optimization, memory management, and robust error-handling watchdogs.

Accomplishments that we're proud of

We successfully proved that a fully functional, enterprise-grade AI agent system can run autonomously on edge devices with minimal operational overhead.

What we learned

Building edge-native AI systems requires balancing API bandwidth with local execution, optimizing lightweight database schemas, and writing resilient self-healing background scripts.

What's next for NEOVRY13 OS

Expanding multi-channel API integrations, improving local model response latencies, and scaling deployment packages for effortless setup by non-technical creators.

Google Technologies Used

  • Gemini 3.5 API: Core intelligence engine for intent evaluation, multi-turn dialogue, and automated lead scoring.
  • Google GenAI SDK: Agent orchestration framework managing context pipelines and tool execution.
  • Google Cloud Run & Cloud SQL: Serverless cloud infrastructure supporting high-availability webhooks and database persistence. Un

Built With

Share this project:

Updates