IRIS
Ecosystem Momentum

RELEASE
CHANGELOG.

Review the structural iterations, feature extensions, and performance upgrades of IRIS.

Version Deprecation & Open Core Transition

🚨 Important Notice: All versions before v1.7.0 are now deprecated and will no longer work. Please use only v1.7.0 and future updates.

In our earlier iterations (pre-v1.1.0), IRIS was developed as a 100% free and open-source project. However, to fund continuous engineering cycles, integrate low-latency SDK solutions (Gemini Live API), and construct advanced tools, IRIS has transitioned to an Open Core model.

Pricing Update: IRIS Pro is available for approximately ₹999 + platform fee.

Sponsorship Tiers & Access:
  • 🟢 $5 a month - IRIS Supporter

    GitHub Sponsor badge, sponsor-only updates, development progress reports, and monthly project updates.

  • ⚡ $15 a month - IRIS Insider

    Everything in Supporter, plus access to the private `iris-insiders` repository, early development previews, architecture discussions, and behind-the-scenes progress.

  • 🚀 $30 a month - IRIS Builder

    Everything in previous tiers, plus access to `iris-labs`, experimental AI agents, prototype systems, technical development logs, and alpha feature testing opportunities.

  • 👁️ $50 a month - IRIS Alpha Access

    Everything in previous tiers, plus access to `iris-alpha`, early alpha builds, upcoming feature previews, private release notes, and priority sponsor feedback consideration.

👑 $250 a month - IRIS MAX

Sponsors will be automatically invited to IRISX-AI/IRIS-MAX.
Everything from previous tiers, plus access to the private IRIS Legacy Repository.

Includes:

  • Full IRIS v1 source archive
  • Legacy voice agent implementation
  • Legacy tool execution system
  • Architecture documentation
  • Development history & design notes
  • Archived releases and examples

Built for developers who want to study how IRIS evolved over time. Note: This tier provides access to the legacy IRIS v1 codebase only. Current proprietary systems, IRIS 2.0, commercial modules, private infrastructure, and future internal development are not included.

SYSTEM CORE: OPERATIONAL
CHRONOLOGY MONITOR
Currently Viewing
v1.7.1Patch
Authentication & Stability Patch

Fixed Lock Screen and Face Unlock, updated cooldown logic, and resolved basic UI bugs.

Ecosystem Versions Index
Return to Documentation
v1.7.1July 13, 2026
Patch Release

Authentication & Stability Patch

Fixed Lock Screen and Face Unlock, updated cooldown logic, and resolved basic UI bugs.

🛠️ Bug Fixes & Improvements

  • •Fixed basic system bugs and applied minor patches.
  • •Fixed Lock Screen behavior and default authentication state.
  • •Updated Face Unlock to load biometric models on-demand, improving performance.
  • •Updated system cooldown mechanisms.
  • •Applied minor UI updates and refinements.
v1.7.0July 2026
Major Release

The Apex Update

Interactive widgets, System Controls, new UI, Free tier limits, new Servers, and full deprecation of all versions prior to 1.7.0.

✨ New Features & Updates

  • •Interactive Widgets: Added new interactive widgets for Quiz (PRO) and Tic Tac Toe (Free).
  • •System Control: Deeper integration for system controls for IRIS.
  • •UI Upgrade: Seamless experience with new Dashboard UI and new Settings UI.
  • •AI State Colors: AI thinking color is now Blue, speaking color is Green, and Standby is White.
  • •Infrastructure: Added new dedicated Server for enhanced stability and control.
  • •Deprecation: All previous versions prior to 1.7.0 are now completely deprecated and no longer function.

🐛 Bug Fixes

  • •Fixed and stabilized Email and WhatsApp integrations.
  • •Fixed Image Generation and Wallpaper Engine.
  • •Fixed Ghost Control issues.
  • •Fixed Camera integration.
  • •Fixed Screen Context capabilities for AI.

💰 Tier Adjustments & Limits

  • •Free Tier Limitations: Introduced a rate limit of 10 conversational turns and 5 tool calls per day. Quota resets every day between 12:00 PM - 2:00 PM.
  • •Wallpaper Limit: Implemented a limit of 5 wallpaper changes per 12 hours.
  • •Vision is Free: The IRIS Vision features (Camera & Screen Context) are now completely free to use for all users.
  • •Tool Access: Updated Free vs Pro tier tools based on the new architecture.
v1.6.3July 7, 2026
Major Release

Advanced Multitasking & Stability Patch

Introduced advanced split-screen app placement for Active Protocols, a premium UI refresh, and critical bug fixes for Face Unlock, Web, and Ghost Voice.

✨ New Features & UI Enhancements

  • •Advanced Active Protocol App Placement: You can now dictate exactly where apps open on your screen (Left, Right, Top, Bottom, or Split-screen) for ultimate multitasking.
  • •Premium UI upgrades and visual refinements across the ecosystem for a better, more polished experience.

🛠️ Bug Fixes & Patches

  • •Fixed Web interaction glitches for smoother browsing automation.
  • •Resolved the Face Unlock glitch to ensure reliable biometric authentication.
  • •Fixed the Wallpaper Engine bug where dynamic backgrounds were failing to apply.
  • •Fixed the Ghost Voice typing glitch causing text to render incorrectly over the screen.
  • •Various other under-the-hood patches and system optimizations.
v1.6.2July 4, 2026
Major Release

Workflow & Personalization Update

Introduced Custom Active Protocols for multi-step automation and the new AI Wallpaper Engine for dynamic desktop backgrounds.

✨ New Features & Capabilities

  • •Added AI Wallpaper Engine to dynamically change PC wallpaper with natural language
  • •Introduced Custom Protocols to build and execute multi-step automated workflows natively in the OS
v1.6.1July 1, 2026
Patch Release

Minor Fixes & Improvements

Fixed a minor bug in the websearch module and resolved issues with Playwright integration.

🛠️ Bug Fixes

  • •Fixed a minor bug in the websearch module
  • •Resolved issues with Playwright integration for improved stability
v1.6.0June 28, 2026
Major Release

The Next Generation IRIS

Introducing Wakeup words, Free & PRO tiers, Chat/Voice toggle, advanced Camera controls, and a fully upgraded Premium UI.

✨ New Features & Capabilities

  • •Added Wakeup word functionality for seamless hands-free activation
  • •Introduced Chat/Voice toggle — use voice or text to interact with IRIS
  • •Advanced Mobile Camera Control: Full control over front/back camera with customized video recording settings
  • •More precise authentication including IP and Location verification

💎 UI & Ecosystem

  • •IRIS now divided into Free and PRO tiers with distinct badges
  • •New quick-access overlay available via Ctrl + Shift + I
  • •Complete UI Upgrade: Simple, premium, and fully animated interface
  • •Updated Settings panel with a new proper build structure

⚡ Performance & Fixes

  • •Quick Build integration resulting in significantly faster app load times
  • •Various underlying bug fixes and system stability improvements
v1.5.1June 25, 2026
Major Release

The Voice Rewrite

A complete overhaul of the communication and voice logic, rebuilding the agent on Google's Gemini Live SDK via native WebRTC.

🎙️ Voice Architecture

  • •Rebuilt entire voice engine using the Gemini Live SDK
  • •Removed legacy WebSocket-heavy communication layer entirely
  • •Fixed speaker feedback loop — mic is now muted while IRIS speaks (software AEC)
  • •Configured VAD sensitivity (START_SENSITIVITY_LOW) to prevent speaker audio being picked up as user speech
  • •Added audioStreamEnd flush after IRIS finishes speaking
  • •Microphone now blocked during AI playback via irisIsSpeaking gate

🧠 Agent & Memory

  • •Improved memory context handling across sessions
  • •Refactored tool execution pipeline for cleaner routing
  • •Vision support improved — push-to-Gemini now works mid-session

🎨 UI

  • •Rebuilt 3D AI orb (AICore.tsx) — particle count reduced from 5,600 → 900 (−84%)
  • •Switched to Fibonacci sphere distribution (deterministic, no rejection loop)
  • •Zero per-frame heap allocations (all THREE.Color, THREE.Vector3 moved to module scope)
  • •Precomputed inverse radius — eliminated per-particle Math.sqrt in render loop
  • •Canvas GL: antialias: false, depth: false, stencil: false, precision: 'lowp' — halved GPU memory
  • •Pixel ratio capped at 1.5x (dpr={Math.min(window.devicePixelRatio, 1.5)})
  • •Torus ring segments reduced 96 → 48, rings reduced 3 → 2
  • •Orbital rings now phase-offset to avoid synchronized pulse

📱 Mobile (ADB)

  • •Fixed executeCameraControl — camera no longer reopens if already running
  • •Fixed front/back lens switching using IMAGE_CAPTURE / VIDEO_CAPTURE intents
  • •Video mode now opens camera in video intent directly (no mode toggle needed)
  • •Added file type validation in polling loop — waits for correct .mp4 / .jpg extension
  • •Extended polling to 15 attempts for long videos
  • •Mux buffer dynamically scaled (duration > 30s → 5s buffer, else 3.5s)
  • •Camera always returns to home screen on failure
v1.3.0June 2, 2026
Minor Release

Stability & Interface Update

Refined window controls, title bar design, and resolved multiple voice playback and desktop workflow bugs.

🛠️ Bug Fixes

  • •Fixed voice interaction glitch causing interruptions during speech playback
  • •Improved voice response consistency for real-time interaction
  • •Resolved multiple internal bugs affecting desktop workflow
  • •Reduced edge-case UI inconsistencies

🎨 UI

  • •Redesigned title bar with cleaner, more polished layout
  • •Improved window controls and visual consistency
  • •Refined hidden internal app behavior and access flow
  • •Enhanced system navigation across modules
  • •General interface polish across the desktop experience

⚙️ Core

  • •Improved performance across multiple interactions
  • •Cleaner system behavior during extended usage
  • •Behind-the-scenes refinements for future IRIS upgrades
v1.2.4May 24, 2026
Patch Release

Infrastructure & Autonomy Update

Secure backend migrations and crucial builder compilation hotfixes to exclude heavy local neural weights.

🚨 Critical Hotfixes

  • •Fixed API blackout caused by accidental deletion of primary Vercel production environment
  • •Provisioned new secure Node.js/Express backend: https://www.irisxai.in
  • •Restored frontend UI: https://www.irisxai.in/
  • •Added Vercel delete protection and strict CORS configuration
  • •Fixed fatal ERR_ELECTRON_BUILDER_CANNOT_EXECUTE (mmap) crash during npm run build:win

📦 Build

  • •Root cause: makensis.exe (32-bit) hitting 2GB limit compressing 4.4GB of local AI models + LanceDB indexes
  • •Fix: set compression: store in electron-builder.yml
  • •Explicitly excluded models/**/*, vectordb/**/*, lancedb/**/* from installer packaging
  • •Windows installer now builds cleanly with lightweight final bundle
v1.2.2May 23, 2026
Patch Release

Telemetry & Stability Patch

Complete hardware telemetry HUD overhaul and automated installers' ghost helper process kills.

🎨 Telemetry HUD (Full Redesign)

  • •Replaced flat metric boxes with glassmorphic hardware telemetry HUD
  • •Each node (CPU, RAM, Temp, OS) has isolated color-coded gradient mesh background (Emerald, Cyan, Orange, Purple)
  • •Added animated CSS cyber-grid patterns per metric card
  • •Added giant faded background iconography (140px) with hover scale animation
  • •Laser-edge hover glow — color-matched per hardware node
  • •Live progress bars in darkened track with backdrop blur

🛠️ Ghost Process Eradication (Auto-Updater Fix)

  • •Root cause: hidden background helper processes blocking installer on restart
  • •Fix: wrapped update execution in setImmediate, stripped all window-all-closed listeners before installer fires
  • •App now aggressively kills own process tree — guarantees 100% update install success

🎙️ Universal Microphone Fix

  • •Fixed microphone silence on specific Windows machines
  • •Expanded Electron session permission handler to approve all Chromium hardware strings: audioCapture, media, microphone, camera
  • •IRIS now has guaranteed audio access regardless of OS build
v1.2.1May 23, 2026
Patch Release

Quantum Core & UI Overhaul

High-frequency AudioWorklet packeter fix to eliminate latency, and new automated update check triggers.

⚡ Zero-Latency Voice Engine (Full Rewrite)

  • •Fixed 8–10 second audio latency — root cause was AudioWorklet flooding WebSocket 60x/sec with micro-packets
  • •Audio now buffered into 4096-frame (250ms) chunks — drops latency to near-zero
  • •Added native VAD and interruption support — speaking while IRIS talks now instantly flushes playback queue
  • •Parallel tool execution via Promise.all — multi-step tool calls now run simultaneously

🎛️ Interactive Auto-Updater

  • •New SYSTEM firmware tab in Command Center
  • •Manual check for updates, real-time patch notes inside app, download progress UI
  • •Disabled forced silent downloads — update now waits for explicit user authorization

🔐 Biometric Vault & Auth

  • •Fixed critical pathing bug (/models → ./models) — Face ID now works in compiled .exe
  • •Added 2.8-second cinematic decryption sequence on Face ID unlock (rotating rings + vault progress bar)
  • •Webcam light turns off instantly on successful Face ID match
  • •New live boot terminal on login screen with simulated OS sequence and glassmorphic OAuth portal

⚡ Low-End Device Optimization

  • •Eliminated new THREE.Color() inside render loop — moved to module scope
  • •Switched array reductions to fast-path math loops — CPU usage reduced by 40%+
  • •WebGL canvas now detects and caps pixel ratio (dpr={[1, 1.5]})
  • •Disabled depth-write for transparent particles — freed GPU bandwidth
v1.1.4May 20, 2026
Patch Release

Firmware Control Patch

Reorganized Command Center UI, introducing user-consented updates logic.

🎛️ System Updates UI

  • •New SYSTEM tab in Command Center for firmware management
  • •Manual update check with in-app patch notes display
  • •Animated download progress UI
  • •Disabled silent forced downloads — all updates require explicit user authorization
  • •Command Center reorganized — System tab now primary landing zone
v1.1.2May 14, 2026
Patch Release

Telemetry & Vault Patch

Real-time socket data graph and native hardware check prompt permissions.

📊 Network Telemetry

  • •Rebuilt Dashboard Neural Uplink panel from scratch
  • •Real-time WSS latency (ping), packet transfer rates (MB/s), routing status (Local/Global)
  • •Animated TX/RX data bars with live fluctuation

🔐 Security

  • •Fixed face-api.js neural network pathing failure in compiled .exe — Face ID now works in production
  • •Added macOS systemPreferences prompts for Camera and Microphone — fixes silent hardware blocks on Apple Silicon and Intel
v1.1.0May 2026
Minor Release

The Agentic Core

LangGraph StateGraph multi-step loops integration and native ADB command execution shell.

🧠 Agent System

  • •Integrated LangGraph StateGraph for autonomous multi-step task orchestration
  • •Tool routing system with dynamic tool registry
  • •Agent interrupt and resume support
  • •Parallel tool execution foundation

📱 ADB Mobile Bridge (Initial)

  • •Wireless Android device connection over TCP/IP
  • •Device telemetry (battery, storage, model, OS version)
  • •Screenshot capture and pull to PC
  • •App open/close, wake/lock/home, swipe and tap
  • •Notification reader (dumpsys notification)
  • •Hardware toggle: WiFi, Bluetooth, Mobile Data, Airplane Mode, Location
  • •File push/pull (/sdcard/Download/)
  • •APK deploy and launch
  • •Camera control (photo + video, front + back lens)

🖥️ Ghost Control (Initial)

  • •Ghost keyboard: type, paste, key press with modifiers
  • •Human-curve mouse movement (Bezier path generation)
  • •Click, double-click, scroll up/down
  • •Volume control via loudness
  • •Screenshot capture to Pictures folder
v1.0.1April 2026
Patch Release

Hotfix

Fixed auto-updater yml configurations and preserved environment keys across local client updates.

🔧 Fixes

  • •Fixed auto-updater latest.yml 404 on fresh installs
  • •Resolved .env key loss on reinstall — keys now preserved across updates
  • •Fixed LanceDB memory graph wipe on reinstall
v1.0.0February 2026
Major Release

Initial Release

The public birth of the IRIS voice-first Neural OS ecosystem layer.

🚀 Core

  • •First public release of IRIS AI for Windows
  • •Electron + Vite + React + TypeScript desktop application
  • •Real-time voice interaction via Gemini 2.5 Flash (WebRTC)
  • •Local LanceDB vector memory with persistent context
  • •Face ID biometric lock via face-api.js
  • •Secure API key vault using Electron safeStorage
  • •Auto-updater via electron-updater + GitHub Releases
  • •NSIS Windows installer with silent update support

🎙️ Voice

  • •Bidirectional real-time audio with Gemini Live API
  • •System instruction injection with user context and memory
  • •Voice Activity Detection (initial implementation)
  • •Persistent conversation memory across sessions

🛠️ Tools (Initial Set)

  • •File system read/write/search
  • •App launcher
  • •Web search via Tavily
  • •WhatsApp messaging (ghost control)
  • •Spotify control (ghost control)
  • •Screenshot capture
  • •Clipboard management
  • •System volume control

🎨 UI

  • •Glassmorphic desktop interface
  • •3D AI orb (Three.js particle system)
  • •Real-time transcript overlay
  • •System telemetry HUD (CPU, RAM, Temp, Network)
  • •Command Center settings panel