Skip to content

IRISX-AI/IRIS-X

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

169 Commits
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 

Repository files navigation


Platform Built With AI Engine License Status


IRIS-X is not an app. It is a neural extension of your Android device.

Speak. It executes. No taps. No menus. Just intent.



What is IRIS-X?

IRIS-X is the mobile evolution of the IRIS AI Desktop OS. Where IRIS gave you an agentic layer over Windows, IRIS-X does the same for Android โ€” but goes further. It physically controls your device, reads your notifications, runs background automations while the screen is off, and communicates through a zero-latency full-duplex voice interface powered by Gemini Live.

It is a closed-source, subscription-based product built for people who want their phone to work for them โ€” not the other way around.

User:  "Hey IRIS, I'm heading to the office."

IRIS:  Silencing phone... โœ“
       Opening Notion... โœ“
       Pulling your 9am calendar brief... โœ“
       Done. You have 3 meetings today. First one in 22 minutes.

Architecture

โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”
โ”‚                    CLIENT โ€” React Native (Expo)          โ”‚
โ”‚  โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ” โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ” โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ” โ”‚
โ”‚  โ”‚  Voice Core  โ”‚ โ”‚  OS Control  โ”‚ โ”‚ Background Tasks โ”‚ โ”‚
โ”‚  โ”‚  Wake word   โ”‚ โ”‚  Intents &   โ”‚ โ”‚  Cron jobs &     โ”‚ โ”‚
โ”‚  โ”‚  STT / TTS   โ”‚ โ”‚  A11y APIs   โ”‚ โ”‚  Geofence watch  โ”‚ โ”‚
โ”‚  โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜ โ”‚
โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
                        โ”‚ HTTPS + JWT
โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ–ผโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”
โ”‚              MIDDLEWARE โ€” Express.js on Vercel            โ”‚
โ”‚   Auth Validator โ†’ Bandwidth Guard โ†’ API Key Injector    โ”‚
โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜
            โ”‚                          โ”‚
โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ–ผโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”  โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ–ผโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”
โ”‚   NEURAL ENGINE       โ”‚  โ”‚   DATABASE โ€” Supabase        โ”‚
โ”‚   Gemini 2.5 Flash    โ”‚  โ”‚   users + subscriptions      โ”‚
โ”‚   Gemini Live API     โ”‚  โ”‚   daily_usage (quota)        โ”‚
โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜  โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜

         โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”
         โ”‚      EDGE โ€” On-Device (Zero Cost)    โ”‚
         โ”‚  Native TTS ยท Wake word ยท Geofence  โ”‚
         โ””โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”˜

Core Features

๐ŸŽ™๏ธ Zero-Latency Neural Voice

Powered by the Gemini Live WebRTC API โ€” real-time, full-duplex voice that lets you interrupt mid-sentence. No loading spinners. No waiting. The conversation flows like talking to a person.

  • Always-on wake word detection (Hey IRIS) running fully on-device
  • Custom voice shortcuts mapped to full action sequences
  • Whisper mode โ€” detects quiet speech and adjusts TTS volume automatically
  • Native TTS for all responses (zero Gemini audio cost)

๐Ÿ“ฑ Deep Android Control

IRIS-X has full dominion over the Android ecosystem via Intents and Accessibility Services.

Capability What it does
App lifecycle Open, close, switch any app by voice
Notification commander Read, reply, dismiss notifications hands-free
Hardware control Brightness, volume, flashlight, Wi-Fi, Bluetooth, DND
App locker Lock any app behind biometric or PIN via voice
Clipboard intelligence Monitors clipboard, auto-suggests actions

โšก Automation Flows

A visual no-code trigger-action builder. Build sequences like:

IF  battery < 20%
โ†’   enable low power mode
โ†’   send WhatsApp to Mom: "my phone is dying"
IF  I arrive at [office location]
โ†’   silence phone
โ†’   open Notion
โ†’   read today's calendar brief aloud

Flows run entirely in the background. You build them once, IRIS-X runs them forever.


๐ŸŒ… Morning Neural Briefing

Every morning at your chosen time, IRIS-X auto-runs a full briefing โ€” spoken aloud:

  • Today's weather
  • Calendar events for the day
  • Unread priority notifications
  • Your saved notes and tasks

Built using cheap text API calls + native TTS. Cost to you: near zero.


๐Ÿ” Security Vault

  • Face App Locker โ€” lock any app behind facial recognition
  • Intruder Selfie โ€” wrong PIN triggers a silent front-camera photo + alert to you
  • Privacy Screen โ€” detects shoulder surfers via front camera and blurs the display

๐Ÿ“ Context-Aware Intelligence

  • GPS Geofencing โ€” trigger automation flows based on location zones
  • Calendar Consciousness โ€” reads your calendar and auto-preps context before meetings
  • Smart Reminders โ€” context-based, not just time-based: "remind me when I open WhatsApp"

Tech Stack

Layer Technology
Mobile client React Native (Expo) + NativeWind
OS control Android Intents + Accessibility Services
Background jobs expo-task-manager + expo-background-fetch
Voice (STT) react-native-voice
Voice (TTS) Native Android TTS engine (free)
Wake word On-device detection
Vision expo-camera + Gemini multimodal
Location expo-location + geofencing
Security middleware Express.js on Vercel
Neural engine Google Gemini 2.5 Flash + Live API
Database & Auth Supabase (PostgreSQL)
Payments Stripe + UPI gateway

Monetization โ€” Neural Bandwidth System

IRIS-X is a subscription-first product designed for profitability from day one.

โ‚น399 โ€“ โ‚น499 / month

To protect margins against API costs, every user receives a daily Neural Bandwidth allocation โ€” a cap on voice minutes and command volume per day. The Vercel middleware enforces this on every request before the Gemini key is injected.

Why this works:

  • Average user won't hit the cap in normal daily usage
  • Power users who want more become upsell targets
  • Server costs stay near zero using the edge architecture below

Cost-saving edge architecture:

  • Background briefings fetch cheap text from Gemini ($4.50 / 1M tokens)
  • Audio is rendered locally using the phone's native TTS engine (completely free)
  • Gemini Live audio output ($0.018/min) is used only for real-time conversation

Subscription Tiers

Feature Free Preview IRIS-X Pro (โ‚น399/mo)
Basic voice commands โœ“ โœ“
Manual app control โœ“ โœ“
10 commands/day โœ“ โ€”
Full Neural Bandwidth โ€” โœ“
Wake word always-on โ€” โœ“
Background automation โ€” โœ“
Automation flow builder โ€” โœ“
Notification commander โ€” โœ“
Morning briefing โ€” โœ“
Intruder Selfie โ€” โœ“
Geofence triggers โ€” โœ“
Priority support โ€” โœ“

Security

IRIS-X is built around a zero-trust architecture. Your API keys are never stored on-device or exposed in the client bundle.

Client  โ†’  sends JWT token + prompt text
           (no API keys ever leave the server)

Middleware  โ†’  validates subscription
            โ†’  checks daily quota
            โ†’  injects Gemini key server-side
            โ†’  forwards to Gemini

Gemini  โ†’  returns response
        โ†’  middleware strips key before forwarding
        โ†’  client receives clean response

All authentication is handled via Supabase with Row Level Security enforced at the database layer.


Roadmap

  • Core architecture designed
  • UI design system (dark neon green theme)
  • Voice core โ€” wake word + STT integration
  • Gemini Live WebRTC streaming
  • Middleware โ€” auth + bandwidth guard on Vercel
  • Supabase schema โ€” users, subscriptions, daily_usage
  • Android Intents โ€” app control layer
  • Accessibility Services โ€” notification reader
  • Automation flow builder UI
  • Geofence trigger system
  • Morning briefing background worker
  • Security vault โ€” face locker + intruder selfie
  • Stripe + UPI payment integration
  • Beta launch (closed)
  • Public launch on Play Store

The IRIS Ecosystem

IRIS-X is part of a broader ecosystem:

Product Platform Status
IRIS AI Windows Desktop โœ… Live โ€” Open Source
IRIS-X Android Mobile ๐Ÿ”จ In Development
IRIS Cloud Web ๐Ÿ”ฎ Planned

About the Architect

Harsh Pandey โ€” AI Systems Engineer

Building agentic OS layers that turn intent into real-world execution. IRIS Desktop crossed 5,000 Instagram followers and 70 GitHub stars in 3 months. IRIS-X is the next chapter.

Instagram GitHub


License

IRIS-X is proprietary software. All rights reserved.

The source code is not open for distribution, modification, or reuse without explicit written permission from the author.

IRIS-X is not an app. It is a neural extension of your Android device.

System initializing.

Made with intent by Harsh Pandey

About

๐Ÿ“ฑ IRIS-MX is a premium, voice-first mobile AI assistant built for high-performance execution. Automate tasks, query web search, and analyze live visual contextโ€”all driven by fast, zero-friction voice commands. Designed as the ultimate pocket productivity tool.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors