▸ 100% FREE & OPEN SOURCE

The AI They Can't See,
Can't Detect,
Can't Stop 👻

invisible • undetectable • unstoppable

An invisible AI overlay that lives on your screen — answers questions, solves problems, analyzes screenshots, and feeds you real-time intelligence. Works during interviews, exams, meetings, or anything on your screen.

⭐ Star on GitHub 📖 Get Started

FREE FOREVER STEALTH MODE upto ~3,000 TOK/S

// 01

What Makes Aura Different

Not another chatbot window. A ghost layer on your screen that nobody else can see.

👻

Invisible Overlay

Screen-capture protected via Win32 API. Hidden from taskbar and Alt+Tab. Ghost mode lets you click through to apps underneath.

UNDETECTABLE

⚡

Stupidly Fast

Powered by Cerebras (upto ~3,000 tok/s) and Groq (upto ~700 tok/s) — the fastest inference engines on Earth. Answers before you finish reading.

SUB-SECOND

📸

Vision AI

Screenshot any problem — code, diagrams, math, schemas. Queue up to 4 screenshots and batch process them all at once.

MULTI-SCREENSHOT

🎤

Voice Intelligence

Live transcription via Deepgram. Aura hears the questions, knows your resume and job description, and generates hyper-personalized responses.

🔄

Multi-Provider AI

Run two AI models simultaneously. Auto-failover across Cerebras, Groq, Gemini, OpenRouter. API key rotation for unlimited throughput.

💸

$0. Forever.

No subscriptions. No credits. No paywalls. 100% free and open source. Bring your own API keys — all providers have free tiers.

// 02

Up and Running in 4 Steps

From clone to invisible AI overlay in under 5 minutes.

Set Up Your Profile

Enter your name, target company, role, resume, and job description. Aura tailors every response to you.

Pick Your AI Models

Choose primary + secondary AI providers. Configure Vision AI for screenshot analysis. Switch between them with hotkeys.

Add Your API Keys

Plug in free API keys from Cerebras, Groq, Gemini, and Deepgram. Add multiple keys per provider for maximum throughput.

Launch & Dominate

Run run.bat for automated setup, or silent_run.vbs for invisible launch with zero terminal windows.

// 03

Stupidly Fast

The two fastest inference engines on Earth. Answers hit your screen before the question finishes loading.

~3,000

tokens per second

Cerebras delivers full, detailed answers in under a second. Groq follows close at upto ~700 tok/s. Both have free tiers.

upto ~3,000 tok/s

upto ~700 tok/s

Vision AI

Live STT

Aura (Cerebras) upto ~3,000 tok/s

Aura (Groq) upto ~700 tok/s

// 04

Stealth Mode: Completely Invisible

Engineered to be undetectable by screen recordings, video calls, and proctoring software.

✓
Screen Capture Protection
Invisible to Zoom, Teams, Meet, Discord, and all screen recording software via WDA_EXCLUDEFROMCAPTURE
✓
Taskbar & Alt+Tab Hidden
Doesn't appear in taskbar, task switcher, or application monitors via WS_EX_TOOLWINDOW
✓
Ghost Mode
Click-through overlay — interact with apps underneath while seeing Aura's content. Toggle with Alt+X
✓
Focus-Free Control
20+ global hotkeys — show, hide, move, scroll, switch AI models. Never click on Aura. Never break focus.
✓
Silent Launch
Start with silent_run.vbs — zero terminal windows, zero system tray icons, zero visible traces.
✓
Variable Transparency
Alt+1 (40%), Alt+2 (70%), Alt+3 (100%) — dial in exactly how invisible you want to be.

// 05

Aura vs. The Rest

Why pay $25/month for slower, weaker, detectable tools?

Feature	👻 Aura	Interview Coder	Parakeet AI	LockedIn AI
💰 Cost	FREE	$25/mo	Credits	Subscription
⚡ Speed	upto ~3,000 tok/s	Slow (GPT)	Sluggish	Laggy
👻 Stealth	Full — Capture proof, ghost mode, taskbar hidden	Half-baked	Partial	❌ Exposed
📸 Vision AI	✅ Multi-screenshot queue + batch processing	❌	❌	❌
🤖 AI Engine	Multi-provider + dual model + key rotation	Single provider	Limited	Basic
📋 Scope	Coding, Design, Behavioral, Aptitude, Quant	Coding only	Limited	Basic