Local Audio Layer
Executes C++/JUCE Core, manages ASIO drivers, VST3 plugin graph, and mixer buses directly in local RAM/CPU.
Available
01
1. Product Architecture Overview
idolLiveAudio is a Windows desktop application built on the Local-First paradigm: offloading heavy real-time DSP, VST3 plugin chains, and AI inferences directly to the user's local CPU/GPU for sub-3ms ultra-low latency with zero dependence on cloud streaming buffers.
Local AI Inference
Embeds local ONNX Runtime to execute lightweight CNN models for Chroma CQT Key & Scale prediction on analysis clips.
AvailableCloud Support Services
ASP.NET Core REST API backend handling identity, license seat validation, My Preset sync, and catalog metadata.
Available
02
2. Real-Time Audio Engine Pipeline
Audio signals from physical input devices flow through the C++/JUCE Engine, execute VST3 vocal chains, and route clean audio output to OBS Studio or Virtual Mic feeds.
Real-Time Audio Processing Pipeline
Local Execution · Sub-3ms Latency
Input
Physical Audio Device
Microphone / Audio Interface
Available
Core Engine
idolLiveAudio Engine
C++ / JUCE Core
Available
DSP / FX
VST3 Processing
Vocal Chains / Effects
Available
Output Feed
Routing & Output
iLA VAD / Virtual Mic / OBS / Rec
In active development
03
3. Local AI Architecture (iLA AutoKey-AI™)
When verified catalog data is missing, iLA AutoKey-AI™ analyzes selected audio clips using a local CNN model running on ONNX Runtime to accurately predict Key & Scale.
iLA AutoKey-AI™ Clip Analysis Pipeline
Local ONNX Inference · On-Demand Clip Analysis
Stage 1
Audio Input / Analysis Clip
Selected audio segment
Available
Stage 2
Chroma CQT Feature
Spectral representation
Available
Stage 3
Lightweight CNN
Proprietary classification model
Available
Stage 4
ONNX Runtime
Local model execution
Available
Stage 5
Key & Scale Prediction
24 musical key classes
Available
04
4. iLA VAD™ & Stream Routing
Our proprietary Windows Virtual Audio Driver and ASIO Bridge provides Virtual Microphone routing, enabling clean processed audio feeds to TikTok LIVE Studio, OBS Studio, and recording software.
05
5. Verified Technology Stack
The verified tools, languages, frameworks, and infrastructure powering idolLiveAudio:
C++
Core Audio Engine & High-Performance DSP
JUCE
Cross-Platform Audio Framework & VST3 Host
VST3
Steinberg Plugin Standard & Vocal Chains
ASIO
Low-Latency Professional Hardware Driver
ONNX Runtime
Cross-Platform Local AI Inference Engine
PyTorch
Offline CNN Model Training & Dataset CQT
ASP.NET Core
Cloud API, Licensing & Web Platform (.NET 10)
PostgreSQL
Relational Database for Catalog & Accounts
Docker
Containerized Backend Cloud Deployment
06
6. Local-First Architectural Benefits
Sub-3ms Latency
Direct hardware buffer processing bypasses network roundtrips, ensuring zero audio delay for live headphone monitoring.
Complete Data Privacy
Live vocal signals stay entirely on your local Windows PC, never streamed or logged to external cloud servers.
Offline Resilience
Audio engine, VST3 chains, and AI models execute offline without requiring active internet connectivity during sessions.
Zero Cloud Streaming Cost
Eliminates high cloud streaming infrastructure costs, allowing iLA Soft to deliver high-performance tools at accessible prices.
07
7. Development Status & Ownership
idolLiveAudio is a proprietary software product owned, designed, and developed by CÔNG TY TNHH ILA SOFT. Currently operating under Commercial Release v1.0 with continuous local audio engine updates.
Developer & Publisher:
CÔNG TY TNHH ILA SOFT (iLA Soft Co., Ltd.)
Corporate Portal ↗
Founder LinkedIn ↗
Company Facebook ↗
Working C++/JUCE Core Audio Engine
Available in v1.0 Windows Desktop App
VST3 Plugin Host & Preset Recall
Available in v1.0 Windows Desktop App
iLA AutoKey-AI™ Local ONNX Runtime
Available On-Demand Clip Analysis
iLA VAD Virtual Audio Driver & ASIO Bridge
In active development & Beta Testing
Cloud License Verification & Preset Sync
Available via ASP.NET Core API