PORTFOLIO DEMONSTRATION — SOUND REQUIRES YOUR PERMISSION
Just A Rather Very Intelligent System
A custom AI system built to understand natural language, speak, remember context, interact with digital environments, control supported computer functions, use tools, coordinate specialised agents and process visual information.
DON'T JUST USE TECHNOLOGY. BUILD THE TECHNOLOGY YOU WISH EXISTED.
“What if I could build my own JARVIS?”
“I decided to find out.”
Instead of only imagining it, Yehia Omar began engineering it. J.A.R.V.I.S. is a real, ongoing engineering project that keeps evolving as new AI, automation, voice, vision and interface capabilities are added.
J.A.R.V.I.S. is designed to route each request through understanding, context and memory before choosing a tool and acting — then verifying the result before it answers.
Hosted models used for heavy reasoning tasks.
On-device models for private, offline-friendly tasks.
Different providers/models selected per task and configuration.
Structured tool invocation instead of plain text replies.
“JARVIS is designed to remember context instead of treating every interaction as a blank conversation.”
CONTEXT PERSISTED ACROSS SESSIONS
A demonstration of the JARVIS voice pipeline. Enable sound to hear the responses rendered through your browser's speech engine.
MCP and structured tool connectivity let JARVIS reach beyond conversation. New capabilities can be attached to the same brain without rebuilding it.
TOOL CATEGORIES SHOWN AS ARCHITECTURE — INDIVIDUAL INTEGRATIONS VARY BY CONFIGURATION
Commands travel from voice, through the JARVIS orchestrator, into a desktop bridge running on the machine — and the result travels back.
DEMONSTRATION ONLY — THIS PAGE CANNOT AND DOES NOT CONTROL YOUR COMPUTER
The vision engine analyses a camera feed; gesture recognition maps hand movement onto interface actions. Both are shown here as demonstrations — your camera is never activated by this page.
Specialised AI personalities coordinated by the JARVIS orchestrator — each with a defined role inside the wider architecture.
Scheduling, reminders, daily flow
Research, summarisation, documents
Conversation and presentation layer
Long-running task execution
Personal context and preferences
Each mode reconfigures the HUD density and what the interface prioritises. Modes are presentation states — they do not add offensive capability.
Standard assistant interface with full tooling.
Beyond conversation, JARVIS is a workspace: 3D objects, parametric design, versioning and project state all live inside the same environment.
A live diagnostics surface keeps every subsystem accountable — from the AI brain to the desktop bridge.
The goal is a system that proactively delivers context — instead of waiting for every command.
Real JARVIS captures are published here as they are supplied. Nothing on this page is a fabricated screenshot — categories without authentic media keep a placeholder.
“I don't start with a template. I start with a question: what should this become?”
JARVIS is an example of that philosophy — taking an ambitious idea, breaking it into systems, and continuously engineering it into something real.
The work is designed from first principles: combining AI, software, automation, voice, vision and interactive interfaces into products engineered around a real problem rather than assembled from templates — and testing emerging technology early enough to use it well.
Y Iconic doesn't only create websites. Y Iconic takes a strange, ambitious, highly specific idea and engineers it into a real digital product.
JARVIS was built for ourselves.
Your custom system could be built for you.
JARVIS started as a personal idea and became a real AI system. Your version doesn't have to look like JARVIS — it can be designed around your business, workflow, ideas and ambitions.
TELL YEHIA WHAT YOU WANT TO BUILD.
JARVIS WAS AN IDEA.
WE BUILT IT.
YOUR IDEA COULD BE NEXT.
YOU IMAGINE IT.
Y ICONIC BUILDS IT.