🎤 VocalCleaner — Dual-Engine AI Audio Enhancer & Mastering Pipeline 🎚️

VocalCleaner is a production-grade, dual-engine AI audio enhancing studio pipeline built with Node.js and Express. It seamlessly chains the surgical noise cleaning power of CleanVoice AI with the premium broadcast-level mastering abilities of Auphonic API, converting raw, noisy recordings into polished, radio-ready studio vocals in seconds.

🚀 Live Demo

VocalCleaner is deployed and live on Render!
👉 Try VocalCleaner Live App Now! 🌐

⚠️ Note: VocalCleaner is hosted on Render's free tier. If the app hasn't been used recently, the first load may take 1–2 minutes while the server wakes up. This is normal — please wait and then refresh the page.

---

🎛️ The Dual-Engine Pipeline (Surgical Approach)

Rather than running a generic double-filtering pass that causes metallic audio distortions, VocalCleaner utilizes an optimized, non-destructive chain:

graph LR
    Raw[Raw MP3/WAV] -->|Upload| CV[CleanVoice AI: Phase 1]
    CV -->|Surgical Cleaning: Noise, Mouth, Breath| Aup[Auphonic API: Phase 2]
    Aup -->|Mastering: Leveler, EQ recovery, Sibilance| Final[Broadcast Studio Output]

Loading

🔬 Phase 1: CleanVoice AI (Surgical Denoise)

• Pure Noise Extraction: Targets and cleans heavy background hums, breath noises, and mouth clicks without affecting the natural voice timbre.
• No Auto-EQ Distortion: Deactivates automated EQ shaping and normalizations in this phase to prevent compression artifacts.

🎚️ Phase 2: Auphonic API (Studio Mastering)

• EQ & Frequency Recovery: Boosts high frequencies (3–8 kHz EQ recovery) to restore natural brightness removed during noise cancellation.
• Adaptive Levelling & De-essing: balances vocal spikes, controls sharp sibilance, and applies broadcast standard -16 LUFS loudness normalization.
• De-noise Gate Deactivated: Kept off during this phase to prevent choppy silence gaps.

---

✨ System Architecture & APIs

• Asynchronous Processing: Accepts audio uploads, writes them to disk to protect RAM, boots up the CleanVoice process, and immediately returns a unique editId to the client. The frontend tracks the job asynchronously via status polling.
• Proxy-Safe Downloader: Features a custom redirect-tracing downloader that securely pulls mastered files from Auphonic S3 buckets, bypassing standard browser CORs policies.
• Responsive Control Dashboard: A premium web UI built with Vanilla JS and CSS3 supporting intensity parameters (Light, Medium, Aggressive) and clean file download actions.

---

🔒 Security Configuration

• Credential Hiding: All cleanvoice and auphonic credentials are securely managed via environment variables (.env).
• Clean Git Footprint: Excludes .env, local node_modules/, local compiled static assets (dist/), and local zip backups (*.zip, *.rar) to prevent security and space overhead on GitHub.

---

🚀 How to Set Up & Run Locally

1. Prerequisites

• Node.js: Ensure Node.js (v18+) is installed.

2. Clone and Configure

git clone https://github.com/oguzemirtopuz/VocalCleaner.git
cd VocalCleaner

Create a .env file in the root folder and add your API credentials:

CLEANVOICE_API_KEY=your_cleanvoice_api_key_here
AUPHONIC_USER=your_auphonic_username_here
AUPHONIC_PASS=your_auphonic_password_here
PORT=3456

3. Install & Start

npm install
npm start

Open http://localhost:3456 in your browser and experience local-first AI audio mastering!

---

👤 Developer Profile

Engineered by Oğuz Emir Topuz.

• Age: 15
• Passions: Football Analyst & Advanced Fullstack Software Developer.
• Connect: My GitHub Portfolio