JSPM

  • ESM via JSPM
  • ES Module Entrypoint
  • Export Map
  • Keywords
  • License
  • Repository URL
  • TypeScript Types
  • README
  • Created
  • Published
  • Downloads 1407
  • Score
    100M100P100Q84015F
  • License MIT

Turn your Obsidian vault into a 3D neural knowledge graph with AI-powered search, memory decay tracking, and MCP integration.

Package Exports

    This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (stellavault) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

    Readme

    Stellavault

    Notes die in folders. Stellavault keeps your knowledge alive.

    Your Obsidian vault is more than files. It's a living network — connections emerge, memories fade, gaps appear. Stellavault turns it into a knowledge system that discovers hidden connections, detects fading memories, finds blind spots, and gives AI agents direct access to everything you know.

    30-Second Setup

    # Install globally
    npm install -g stellavault
    
    # Interactive setup wizard
    stellavault init
    
    # Or manually: index your vault + launch 3D graph
    stellavault index /path/to/obsidian/vault
    stellavault graph

    That's it. Open http://localhost:5173 to see your knowledge come alive in 3D.

    Give Your AI Agent Memory

    claude mcp add stellavault -- stellavault serve

    Now Claude Code can search your notes, track your decisions, tell you what's fading, and give you a daily knowledge briefing — using 17+ MCP tools.

    What It Does

    Understands — Vectorizes every note locally (no API keys needed). Hybrid search combines BM25 + cosine similarity + RRF fusion.

    Remembers — FSRS spaced repetition tracks what you're forgetting. stellavault decay shows what's fading. stellavault review runs a review session.

    Discovers — Gap detector finds missing connections. Duplicate finder catches redundancy. Contradiction detector spots conflicting statements. Learning path tells you what to study next.

    Visualizes — Interactive 3D neural graph with constellation view, timeline slider, type filters, health dashboard, and keyboard navigation.

    Connects — Federation protocol links your vault to other Stellavault nodes via P2P. Search across the network without sharing your raw text. Multiverse UI shows your universe alongside connected peers.

    Key Features

    Category Features
    Search Hybrid BM25+Vector+RRF, semantic search, search history
    Intelligence FSRS decay, gap detection, duplicates, contradictions, predictive gaps, learning paths
    Visualization 3D graph, constellation view, timeline, type filter, health dashboard, embed widget
    AI Integration 17+ MCP tools, custom MCP tool builder (YAML), agentic graph construction
    Federation P2P search, multiverse UI, reputation system, trust scoring, differential privacy
    Security E2E cloud sync (AES-256-GCM), team auth (RBAC), sharing controls, SSRF/XSS protection
    Platform Web dashboard, PWA, voice capture (Whisper), Notion sync, multi-vault, i18n (en/ko/ja/zh)
    Extensibility Plugin SDK, webhook system, Knowledge Pack marketplace

    CLI (28+ commands)

    # Core
    stellavault init                  # Interactive setup wizard
    stellavault index <vault-path>    # Vectorize your vault
    stellavault search <query>        # Semantic search
    stellavault graph                 # Launch 3D graph
    stellavault serve                 # MCP server for AI agents
    stellavault status                # Index stats
    
    # Intelligence
    stellavault decay                 # What's fading?
    stellavault learn                 # AI learning path
    stellavault brief                 # Daily knowledge briefing
    stellavault digest                # Weekly report
    stellavault review                # FSRS review session
    stellavault gaps                  # Knowledge gap detection
    stellavault duplicates            # Duplicate detection
    stellavault contradictions        # Contradiction detection
    
    # Federation
    stellavault federate join         # Join P2P network
    stellavault federate status       # Node identity
    
    # Multi-Vault
    stellavault vault add <id> <path> # Register a vault
    stellavault vault search-all <q>  # Search across all vaults
    
    # Cloud
    stellavault cloud sync            # E2E encrypted backup
    stellavault cloud restore         # Restore from cloud
    
    # More
    stellavault clip <url>            # Web/YouTube clipper
    stellavault sync                  # Notion → Obsidian
    stellavault capture <audio>       # Voice → knowledge (Whisper)
    stellavault card                  # SVG profile card
    stellavault pack <cmd>            # Knowledge Pack management

    Use short alias sv for all commands: sv graph, sv decay, sv learn

    MCP Tools (17+)

    Tool What it does
    search Hybrid search (BM25 + vector + RRF)
    get-document Full document with metadata
    get-related Semantically similar documents
    list-topics Topic cloud
    get-decay-status Memory decay report
    get-morning-brief Daily knowledge briefing
    get-learning-path AI learning recommendations
    federated-search Search across P2P network
    create-knowledge-node AI creates notes during conversation
    create-knowledge-link AI creates links between notes
    log-decision / find-decisions Technical decision journal
    create-snapshot / load-snapshot Context snapshots
    generate-claude-md Auto-generate CLAUDE.md
    export JSON/CSV export

    Federation — P2P Knowledge Network

    Each Stellavault is a node. Connect to discover knowledge across the network — without sharing your raw text.

    stellavault federate join --name "my-node"
    # Interactive mode:
    federation> search kubernetes deployment
    #  87% "K8s Rolling Update Guide" [peer-alice]
    #     Kubernetes에서 무중단 배포를 위해...
    federation> peers
    #  alice (1,209 docs) [trust: 88]
    #  bob (47 docs) [trust: 72]

    Privacy: Only titles + similarity scores + 50-char masked snippets cross the network. Never raw text. Differential privacy noise on embeddings. Sharing controls per tag/folder.

    Trust: Web of Trust + automatic reputation scoring (consistency, consensus, helpfulness).

    Architecture

    stellavault/
    ├── packages/
    │   ├── core/       Search, MCP, API, Intelligence, Federation, Cloud, Team, i18n
    │   ├── cli/        28+ commands
    │   ├── graph/      3D visualization + Multiverse UI (React Three Fiber)
    │   └── sync/       Notion ↔ Obsidian sync

    Tech Stack

    Layer Tech
    Runtime Node.js 20+ (ESM, TypeScript)
    Vector Store SQLite-vec (local, no server)
    Embedding all-MiniLM-L6-v2 (local, no API key)
    Search BM25 + Cosine + RRF Fusion
    3D React Three Fiber + Three.js + Zustand
    Memory FSRS (Free Spaced Repetition Scheduler)
    P2P Hyperswarm (NAT traversal + DHT)
    AI MCP (Model Context Protocol)
    Security AES-256-GCM, HMAC-SHA256, Differential Privacy

    Configuration

    {
      "vaultPath": "/path/to/obsidian/vault",
      "dbPath": "~/.stellavault/index.db",
      "embedding": { "model": "local", "localModel": "all-MiniLM-L6-v2" },
      "search": { "defaultLimit": 10, "rrfK": 60 },
      "mcp": { "mode": "stdio", "port": 3333 }
    }

    License

    MIT