Skip to content

KittLabsAI/KittyNest

Repository files navigation

KittyNest

KittyNest UI Concept

Local-first memory tracker for Claude Code & Codex

简体中文


KittyNest is a local-first, privacy-centric macOS desktop application that helps you track, organize, and make sense of your AI coding sessions from Claude Code and Codex. All your data stays on your machine — sessions are indexed locally in SQLite, and insights are stored as readable Markdown files.

✨ Features

  • 🔍 Session Discovery — Automatically scan and import sessions from Claude Code (~/.claude) and Codex (~/.codex)
  • 📁 Project Tracking — Group sessions by working directory, review project health, and generate project summaries
  • 📝 Task Management — Organize sessions into tasks with statuses (Discussing → Developing → Done)
  • 🧠 Memory System — Three-tier cascading memory (Session → Project → System) with entity-relationship graph
  • 🤖 LLM-Powered Analysis — Analyze sessions and projects using your own API keys (OpenAI-compatible, Anthropic-compatible, and more)
  • 📊 Dashboard — At-a-glance view of active projects, open tasks, recent sessions, and memory status
  • 🔒 Local-First & Private — All data stored in ~/.kittynest; nothing leaves your machine unless you explicitly send it to your configured LLM provider

🏗️ Architecture

┌─────────────────────────────────────────────────────────────┐
│  React Frontend (Vite + TypeScript)                         │
│  Dashboard · Projects · Tasks · Sessions · Memories · Settings│
└──────────────────────┬──────────────────────────────────────┘
                       │ invoke
┌──────────────────────▼──────────────────────────────────────┐
│  Tauri 2 Shell                                              │
│  macOS Window · Menu · File System Permissions · Lifecycle  │
└──────────────────────┬──────────────────────────────────────┘
                       │
┌──────────────────────▼──────────────────────────────────────┐
│  Rust Backend                                               │
│  Commands · Adapters · Jobs · LLM Client · Markdown Store   │
└──────────────────────┬──────────────────────────────────────┘
        ┌──────────────┴──────────────┐
        ▼                             ▼
┌───────────────┐           ┌─────────────────────┐
│  SQLite Index │           │  Markdown Store     │
│  (local cache)│           │  ~/.kittynest/      │
└───────────────┘           └─────────────────────┘

🚀 Getting Started

Prerequisites

  • Node.js 18+
  • Rust 1.70+
  • macOS (primary target platform)

Development

# Install frontend dependencies
npm install

# Run in development mode (opens Tauri window)
npm run tauri dev

Build

# Build production bundle (.app and .dmg)
npm run tauri build

The build artifacts will be located in src-tauri/target/release/bundle/.

📂 Data Storage

All application data is stored locally under ~/.kittynest/:

~/.kittynest/
├── config.toml              # LLM provider settings
├── kittynest.sqlite         # SQLite index database
├── kittynest_graph.db       # Entity-relationship graph database
├── projects/
│   └── <project_slug>/
│       ├── info.md          # Project summary
│       ├── progress.md      # Project progress
│       └── <task_slug>/
│           ├── summary.md
│           ├── user_prompt.md
│           └── <session>.md
└── memories/
    ├── sessions/
    │   └── <session_slug>/
    │       └── memory.md
    ├── projects/
    │   └── <project_name>.md
    └── system/
        └── memory.md

⚙️ Supported LLM Providers

KittyNest supports a wide range of LLM providers via presets:

  • OpenRouter
  • DeepSeek
  • Zhipu GLM
  • Bailian
  • Kimi (Moonshot)
  • StepFun
  • MiniMax
  • DouBao (Seed)
  • ModelScope
  • Ollama (local)
  • OpenAI-compatible APIs

You can configure your provider in the Settings page.

🛠️ Tech Stack

Layer Technology
Frontend React 18, TypeScript, Vite
Desktop Shell Tauri 2
Backend Rust
Database SQLite (rusqlite)
Graph DB CozoDB (SQLite-backed) / Pure SQLite fallback
UI Components Lucide React, XYFlow
Styling Custom CSS

🗺️ Roadmap

  • Project skeleton & configuration
  • Claude Code / Codex source adapters
  • Project tracking & manual review
  • Historical session batch analysis
  • Incremental session scanning
  • Memory module (Session / Project / System)
  • Task creation & management
  • Settings page & LLM configuration
  • macOS desktop integration
  • Graph query UI
  • Memory versioning & rollback
  • Auto-updater
  • Apple Silicon optimization

🤝 Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.

📄 License

MIT


Built with ❤️ for local-first AI session management

About

Local-first memory tracker for Claude Code & Codex

Topics

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors