👻 PhantomEar

Always Listening. Never Seen.

The privacy-first AI meeting assistant that lives on your machine

Download • Features • How It Works • Tech Stack • Development

---

🎯 The Problem

Traditional meeting assistants expose your presence:

• 🤖 Bot joins the call → Everyone knows you're recording
• ☁️ Audio uploaded to cloud → Your data leaves your control
• 🔒 Admin install required → IT blocks the tool
• 💰 Monthly subscriptions → Costs add up

✨ The PhantomEar Difference

🔒 100% Local Processing Audio never leaves your machine. Transcription runs locally using Whisper. Your meetings, your data, your control.	👻 Zero Meeting Footprint No bots joining calls. No recording banners. No participant notifications. Completely invisible.
🧠 AI-Powered Intelligence Ask questions about any meeting. Get instant summaries. Search across all your conversations with semantic search.	🏢 Enterprise Ready User-level installation. No admin rights. No IT tickets. Works on restricted corporate machines.

---

📥 Download

🚀 Download Latest Release

Platform	Architecture	Download
macOS	Apple Silicon (M1/M2/M3)	.dmg
Windows	x64	.exe / .msi

📋 Installation Notes

macOS

Since PhantomEar is not yet notarized with Apple, you may see a "damaged" warning. Fix it with:

xattr -cr /Applications/PhantomEar.app

Windows

• If .msi is blocked by corporate policy, use the .exe installer
• If you see MSVCP140.dll error, install Visual C++ Runtime

---

🎬 Features

Real-Time Transcription Live speech-to-text powered by Whisper	AI Q&A Ask questions about any meeting	Auto Summaries Key points, decisions & action items
Semantic Search Find anything across all meetings	Persistent Memory SQLite with full-text search	100% Private Everything stays on your device

Feature Highlights

Feature	Description
🎙️ Multi-Model ASR	Whisper (tiny → large) or Parakeet CTC models
🤖 LLM Flexibility	OpenAI API or Ollama (fully local)
⏸️ Pause/Resume	Pause transcription without stopping the session
📝 Transcript Editing	Edit segments, fix transcription errors
🗣️ Speaker Labels	Manually assign speakers to segments
📊 Timeline View	Visual timeline with quick navigation
🔍 Quick Search	Cmd+K to search across all meetings
🖥️ System Tray	Minimize to tray, start/stop from menu
📋 Export	Copy as Markdown or plain text

---

🔬 How It Works

┌─────────────────────────────────────────────────────────────────────────┐
│                           YOUR MACHINE                                   │
│  ┌─────────────┐    ┌─────────────┐    ┌─────────────┐    ┌──────────┐ │
│  │  Microphone │───▶│   Whisper   │───▶│   SQLite    │───▶│    UI    │ │
│  │   (cpal)    │    │  (local)    │    │  + FTS5     │    │ (Svelte) │ │
│  └─────────────┘    └─────────────┘    └─────────────┘    └──────────┘ │
│                            │                  │                         │
│                            ▼                  ▼                         │
│                     ┌─────────────┐    ┌─────────────┐                  │
│                     │ Embeddings  │    │  Semantic   │                  │
│                     │ (BGE-small) │───▶│   Search    │                  │
│                     └─────────────┘    └─────────────┘                  │
│                                               │                         │
└───────────────────────────────────────────────│─────────────────────────┘
                                                │
                          ┌─────────────────────┴─────────────────────┐
                          │         OPTIONAL: LLM Context              │
                          │    (Only curated text, never raw audio)    │
                          └─────────────────────┬─────────────────────┘
                                                │
                                    ┌───────────┴───────────┐
                                    ▼                       ▼
                             ┌──────────┐            ┌──────────┐
                             │  Ollama  │            │  OpenAI  │
                             │ (local)  │            │  (API)   │
                             └──────────┘            └──────────┘

Privacy by Design

Data Type	Location	Sent to Cloud?
🎤 Raw Audio	RAM only (never saved)	❌ Never
📝 Transcripts	Local SQLite	❌ Never
🧠 Embeddings	Local SQLite	❌ Never
💬 LLM Context	Selected text only	⚠️ Optional*

*Only if using OpenAI. Use Ollama for 100% local operation.

---

🏗️ Architecture

phantom-ear/
├── src/                          # Svelte 5 Frontend
│   ├── routes/+page.svelte       # Main application
│   ├── lib/
│   │   ├── components/           # UI components
│   │   ├── stores/               # State management
│   │   └── utils/                # Utilities
│   └── app.css                   # TailwindCSS v4
│
├── src-tauri/                    # Rust Backend
│   └── src/
│       ├── lib.rs                # Tauri entry point
│       ├── commands.rs           # IPC handlers
│       ├── audio/                # Audio capture (cpal)
│       ├── asr/                  # Whisper integration
│       ├── transcription/        # Real-time pipeline
│       ├── embeddings/           # BGE-small vectors
│       ├── storage/              # SQLite + FTS5
│       └── llm/                  # OpenAI/Ollama clients

Tech Stack

Rust

Tauri 2.0

Svelte 5

TailwindCSS

Layer	Technology	Purpose
Runtime	Tauri 2.0	Lightweight native wrapper
Backend	Rust	Performance-critical processing
Frontend	Svelte 5	Reactive UI with runes
Styling	TailwindCSS v4	Dark theme, glassmorphism
ASR	whisper-rs	Local speech recognition
Embeddings	ONNX Runtime	BGE-small vectors
Database	SQLite + FTS5	Full-text search
Audio	cpal	Cross-platform capture

---

🛠️ Development

Prerequisites

# Node.js 18+
node --version  # v18.0.0+

# Rust (latest stable)
rustc --version  # 1.75.0+

# CMake (for whisper.cpp)
cmake --version  # 3.20+

Install CMake

# macOS
brew install cmake

# Windows
choco install cmake

# Linux
sudo apt install cmake

Quick Start

# Clone
git clone https://github.com/Phantom-Ear/phantom-ear.git
cd phantom-ear

# Install dependencies
npm install

# Run in development mode
npm run tauri dev

# Build for production
npm run tauri build

---

🗺️ Roadmap

Status	Feature	Target
✅	Real-time transcription	v0.1.0
✅	Meeting persistence	v0.1.0
✅	AI Q&A (RAG)	v0.1.0
✅	Semantic search	v0.1.0
✅	System tray	v0.2.0
✅	Transcript editing	v0.2.0
✅	Speaker labels	v0.2.0
🔄	Audio device selection	v0.3.0
📋	SRT subtitle export	v0.3.0
📋	Auto-meeting detection	v0.4.0
📋	Light theme	v0.4.0
📋	Auto-titling with AI	v0.5.0

---

🤝 Contributing

Contributions are welcome! Please read our contributing guidelines before submitting a PR.

# Fork the repo
# Create your feature branch
git checkout -b feature/amazing-feature

# Commit your changes
git commit -m "Add amazing feature"

# Push to the branch
git push origin feature/amazing-feature

# Open a Pull Request

---

📄 License

MIT License - see LICENSE for details.

---

Built with 🖤 for privacy advocates

⬆ Back to Top

_{PhantomEar is not affiliated with any meeting platform. Use responsibly and in compliance with applicable laws.}