Skip to content

officialnullobjectweb/dead-reckoning

Repository files navigation

Dead Reckoning — Memory OS

Your life's memory. Locally owned. AI-powered. Zero cloud. Zero cost.

A powerful desktop application that indexes your files, lets you search across them in natural language, and chat with your data using local or cloud AI models — all with zero cloud dependency.

License Platform Node


✨ Features

🔍 Full-Text Search

  • Instant search across all indexed files with fuzzy matching
  • Filter by file type (Documents, Notes, Text, Code)
  • Date-based filtering and multiple sort options
  • Command palette with ⌘K keyboard shortcut

🤖 AI Chat

  • Chat with your files using natural language
  • 6 AI providers supported: Ollama (local), OpenAI, Anthropic Claude, Google Gemini, Groq, Offline
  • 50+ models available across all providers including Llama 3.3, GPT-4o, Claude 3.7 Sonnet, Gemini 2.0, DeepSeek R1, and more
  • Real-time streaming responses with markdown rendering
  • File context injection for accurate, grounded answers
  • Persistent chat sessions with history

📂 File Indexing

  • Automatic file watching and indexing with Chokidar
  • Supports 11 file types: .md, .txt, .pdf, .docx, .doc, .rtf, .csv, .json, .js, .ts, .py
  • Real-time indexing progress with live percentage bar
  • Auto re-indexing on file changes

📊 Dashboard

  • At-a-glance stats: total files, words, folders, last indexed
  • Beautiful activity chart showing indexing trends
  • Recently added files with quick access to chat

📅 Timeline

  • Chronological file view grouped by year and month
  • Collapsible month sections with file counts
  • Color-coded file type indicators

🗂️ Vault Management

  • Add/remove watched folders with native file picker
  • Folder cards with file counts, status indicators, and last indexed time
  • One-click re-indexing per folder

⚙️ Settings

  • Provider selection with visual cards
  • Beautiful model dropdown with model name, size, and organization
  • Live model syncing from Ollama, OpenAI, and Groq APIs
  • API key management with show/hide toggle
  • Connection testing for local providers
  • Indexing preferences (auto re-index, hidden files, extension selection)

🎨 Design

  • Premium dark/light theme with smooth transitions
  • Glass morphism effects and gradient accents
  • Smooth Framer Motion animations throughout
  • Responsive and polished UI

🛠️ Tech Stack

Layer Technology
Desktop Electron 41
Frontend React 19 + Vite 5
Routing React Router DOM 7 (HashRouter)
Styling Tailwind CSS 4 + CSS Variables
Animations Framer Motion
State Zustand
Database SQLite (better-sqlite3)
Search MiniSearch (in-memory full-text)
File Watcher Chokidar
Charts Recharts
Markdown React Markdown
Notifications Sonner
Icons Lucide React
UI Primitives Radix UI

📦 Installation

Prerequisites

  • Node.js 18 or higher
  • npm or yarn
  • Ollama (optional, for local AI) — download from ollama.ai

Step 1: Clone the Repository

git clone https://github.com/yourusername/dead-reckoning.git
cd dead-reckoning

Step 2: Install Dependencies

npm install

Step 3: Run in Development Mode

npm run dev

This launches both the Vite dev server and Electron app simultaneously.

Step 4: Build for Production

npm run build

This creates platform-specific installers (DMG for macOS, NSIS for Windows, AppImage/deb for Linux) in the dist/ directory.


🚀 Quick Start

  1. Launch the app — You'll see the dashboard with stats and recent files
  2. Add folders — Go to Vault → Click Add Folder → Select a directory to index
  3. Search — Press ⌘K or go to Search to find files instantly
  4. Chat — Go to Chat → Start a new session → Ask questions about your files
  5. Configure AI — Go to Settings → Choose your AI provider → Select a model → Save

🤖 Supported AI Providers

Provider Cost Models Setup
Ollama Free Llama 3.2, Mistral, Phi-4, Gemma 2, Qwen 2.5, DeepSeek R1, and more Install Ollama, pull models
Groq Free tier Llama 3.3 70B, Mixtral 8x7B, Gemma 2 9B, DeepSeek R1 Create API key
Google Gemini Free tier Gemini 2.0 Flash, 1.5 Pro, 1.5 Flash Create API key
OpenAI Pay-per-use GPT-4o, o1, o3 Mini, GPT-3.5 Turbo Create API key
Anthropic Claude Pay-per-use Claude 3.7 Sonnet, 3.5 Sonnet, 3.5 Haiku, Opus Create API key
Offline Free None (search & browse only) No setup needed

📁 Project Structure

dead-reckoning/
├── electron/                  # Electron main process
│   ├── main.js               # App window, IPC handlers, service init
│   ├── preload.js            # Context bridge (window.dr API)
│   └── services/
│       ├── db.js             # SQLite database layer
│       ├── indexer.js        # File watcher & parser
│       ├── search.js         # MiniSearch full-text engine
│       └── ollama.js         # Multi-provider AI service
├── src/                      # React frontend
│   ├── main.jsx              # Entry point
│   ├── App.jsx               # Root component & routing
│   ├── index.css             # Global styles & theme
│   ├── components/
│   │   ├── Sidebar.jsx       # Navigation sidebar
│   │   └── SearchBar.jsx     # Command palette (⌘K)
│   ├── pages/
│   │   ├── Home.jsx          # Dashboard
│   │   ├── Search.jsx        # Full-page search
│   │   ├── Chat.jsx          # AI chat interface
│   │   ├── Timeline.jsx      # Chronological file view
│   │   ├── Vault.jsx         # Folder management
│   │   └── Settings.jsx      # AI & indexing config
│   ├── store/
│   │   ├── useAppStore.js    # Global app state
│   │   └── useChatStore.js   # Chat session state
│   └── context/
│       └── ThemeContext.jsx  # Light/dark theme provider
├── package.json
├── vite.config.js
└── electron-builder.config.js

🔒 Privacy & Security

  • All data stays on your device — files are never uploaded
  • Local-first architecture — SQLite database stored in your user data directory
  • No cloud dependency — works entirely offline with Ollama
  • Your API keys stay local — stored encrypted in your user data directory
  • Open source — fully auditable codebase

🎯 Benefits

  • Own your data — No third-party servers, no data collection, no tracking
  • Free forever — With Ollama, there are zero ongoing costs
  • Fast & responsive — Local SQLite + in-memory search means instant results
  • Extensible — Support for multiple AI providers and file types
  • Beautiful — Premium UI with dark/light themes and smooth animations
  • Cross-platform — Works on macOS, Windows, and Linux

📝 License

MIT License — see LICENSE for details.


🙏 Acknowledgments

Built with love using open-source tools and frameworks. Special thanks to the Electron, React, and Ollama communities.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors