JDeduplix is a cutting-edge deduplication system that leverages artificial intelligence and decentralized technologies to provide intelligent, accurate, and efficient data deduplication across multiple data types. Built with Rust for performance and safety, it combines advanced AI models with peer-to-peer networking and CRDT-based version control to create a powerful, flexible, and scalable solution.
- Multi-Modal AI Deduplication: Specialized AI models for text, JSON, images, and binary data
- Pluggable P2P Architecture: Support for both LibP2P and Holepunch protocols out of the box
- Distributed Processing: Scale horizontally with distributed worker nodes
- Smart Conflict Resolution: ML-powered conflict detection with human-in-the-loop capabilities
- High-Performance Vector Search: FAISS HNSW for efficient similarity matching
- Modern UI: Tauri + Vue frontend with real-time visualizations
- Flexible Storage: Multiple backend options including decentralized storage
- CRDT-Based Version Control: Loro-powered versioning with automatic conflict resolution
- Rich Data Types: Support for complex data structures with semantic merging
- Enterprise Data Management: Eliminate redundant data across large organizations
- Content Distribution: Efficient content sharing with decentralized storage
- Digital Asset Management: Smart deduplication of media files and assets
- Database Optimization: Reduce storage costs with intelligent data deduplication
- Collaborative Environments: P2P sharing with version control and conflict resolution
- Edge Computing: Distributed processing with local and network storage options
- Collaborative Deduplication: Real-time collaboration with automatic conflict resolution
- Version-Aware Storage: Track and manage data versions with semantic understanding
- Smart Classification: Automatic content type detection and routing
- Specialized Engines:
- ML Text Deduper for documents and code
- GNN JSON Deduper for structured data
- CNN Image Deduper for visual content
- Deep Learning Binary Deduper for raw data
- Intelligent Matching:
- Fuzzy matching with confidence scoring
- Perceptual hashing for media files
- Graph-based structural similarity
- Vector embedding comparison
- Pluggable P2P Layer:
- LibP2P with IPFS storage
- Holepunch with Hypercore support
- Protocol-agnostic adapter interface
- Cross-protocol communication
- Distributed Processing:
- Horizontal scaling with worker nodes
- Load balancing and task distribution
- Fault tolerance and recovery
- Real-time progress monitoring
- Multi-Backend Support:
- High-performance Sled DB
- IPFS for decentralized storage
- Hypercore for append-only data
- Vector databases (Weaviate/Pinecone)
- Intelligent Caching:
- Smart cache with vector indexing
- Content-aware caching strategies
- Distributed cache coordination
- Automatic cache invalidation
- Loro Integration:
- High-performance CRDT implementation
- Rich data type support (Text, List, Map, Tree)
- Automatic semantic merging
- Branch and fork support
- Version Management:
- Complete DAG-based history tracking
- Branch-based workflow
- Automatic conflict resolution
- Manual conflict resolution when needed
- Collaborative Features:
- Real-time multi-user editing
- Branch-level permissions
- Version diffing and comparison
- Semantic merge previews
- Performance Optimizations:
- Fast document loading (sub-millisecond)
- Shallow snapshot support
- Block-level storage
- Selective history loading
- ML Conflict Resolution:
- Confidence-weighted decisions
- Automated threshold tuning
- Human review integration
- Learning from feedback
- Version Control:
- Full history tracking
- Change auditing
- Rollback capabilities
- Branch management
- API Integration:
- REST and WebSocket APIs
- Client SDKs
- Real-time events
- Comprehensive documentation
- Framework: Tauri + Vue.js
- UI Components: Custom-built with dark theme
- Visualizations: Real-time graphs and heatmaps
- P2P Controls: Protocol selection and network management
- Core: Rust for performance and safety
- AI Framework: Custom ML pipeline
- Vector Search: FAISS HNSW
- P2P: LibP2P and Holepunch implementations
- Version Control: Loro CRDT (v1.0+)
- Rich CRDT types
- High-performance DAG
- Semantic merging
- Block-level storage
- Local:
- Sled DB for high performance
- Loro block storage for versions
- Distributed:
- IPFS and Hypercore
- CRDT-aware replication
- Vector: Weaviate/Pinecone integration
- Version History: Loro DAG-based tracking
- Text CRDT: Smart merging of formatted text
- List CRDT: Ordered collections with move support
- Map CRDT: Key-value structures
- Tree CRDT: Hierarchical data with move operations
- Nested Types: Complex JSON-like structures
- Branch Operations:
- Create/switch branches
- Fork for independent work
- Merge with automatic conflict resolution
- Rebase/squash support (WIP)
- History Navigation:
- Checkout specific versions
- Browse version DAG
- Compare versions
- Selective loading
- Real-time Features:
- Multi-user editing
- Automatic merging
- Conflict prevention
- Change visualization
- Integration:
- P2P synchronization
- Distributed storage
- External VCS bridges
- API access
%%{init: {'theme': 'dark', 'themeVariables': { 'darkMode': true }}}%%
flowchart TD
classDef primary fill:#42b883,stroke:#35495e,stroke-width:2px,color:#1a1c1d
classDef secondary fill:#3eaf7c,stroke:#2c3e50,stroke-width:2px,color:#1a1c1d
classDef action fill:#4fc08d,stroke:#2c3e50,stroke-width:2px,color:#1a1c1d
classDef api fill:#ff9f43,stroke:#e67e22,stroke-width:2px,color:#1a1c1d
classDef storage fill:#00b894,stroke:#00cec9,stroke-width:2px,color:#1a1c1d
classDef dist fill:#6c5ce7,stroke:#a29bfe,stroke-width:2px,color:#1a1c1d
classDef backend fill:#0984e3,stroke:#74b9ff,stroke-width:2px,color:#1a1c1d
classDef p2p fill:#00b894,stroke:#00cec9,stroke-width:2px,color:#1a1c1d
classDef version fill:#fd79a8,stroke:#e84393,stroke-width:2px,color:#1a1c1d
subgraph UI[User Interface Layer]
direction TB
FE[Frontend App]:::primary
SDK[Client SDK]:::secondary
API[API Gateway]:::api
end
subgraph BE[Backend Layer]
direction TB
AI[AI Processing]:::backend
IDX[Smart Indexing]:::backend
DIST[Distributed Coordinator]:::dist
end
subgraph ST[Storage Layer]
direction TB
DB[(Local DB)]:::storage
FS[File System]:::storage
VDB[(Vector DB)]:::storage
end
subgraph P2P[P2P Network]
direction TB
LIBP2P[LibP2P Stack]:::p2p
HOLE[Holepunch Stack]:::p2p
DHT[DHT]:::p2p
IPFS[IPFS]:::p2p
end
subgraph VCS[Version Control Service]
direction TB
CRDT[Loro CRDT]:::version
BRANCH[Branch Manager]:::version
MERGE[Merge Handler]:::version
DAG[History DAG]:::version
BLOCK[Block Storage]:::version
CRDT --> BRANCH
CRDT --> MERGE
BRANCH --> DAG
MERGE --> DAG
DAG --> BLOCK
end
FE --> SDK
SDK --> API
API --> AI
API --> IDX
API --> DIST
API --> VCS
AI --> DB
AI --> VDB
IDX --> VDB
IDX --> FS
DIST --> P2P
VCS --> P2P
LIBP2P --> DHT
LIBP2P --> IPFS
HOLE --> DHT
style UI fill:#1a1c1d,stroke:#42b883,stroke-width:2px
style BE fill:#1a1c1d,stroke:#0984e3,stroke-width:2px
style ST fill:#1a1c1d,stroke:#00b894,stroke-width:2px
style P2P fill:#1a1c1d,stroke:#00b894,stroke-width:2px
style VCS fill:#1a1c1d,stroke:#fd79a8,stroke-width:2px
%%{init: {'theme': 'dark', 'themeVariables': { 'darkMode': true }}}%%
flowchart TD
classDef primary fill:#42b883,stroke:#35495e,stroke-width:2px,color:#1a1c1d
classDef secondary fill:#3eaf7c,stroke:#2c3e50,stroke-width:2px,color:#1a1c1d
classDef action fill:#4fc08d,stroke:#2c3e50,stroke-width:2px,color:#1a1c1d
classDef api fill:#ff9f43,stroke:#e67e22,stroke-width:2px,color:#1a1c1d
classDef p2p fill:#00b894,stroke:#00cec9,stroke-width:2px,color:#1a1c1d
classDef version fill:#fd79a8,stroke:#e84393,stroke-width:2px,color:#1a1c1d
subgraph UI[User Interface]
direction TB
A[User Dashboard]:::primary
B[Upload Interface]:::primary
C[Results View]:::primary
D[Smart Classifier UI]:::secondary
E[Manual Resolution]:::secondary
F[AI Confidence Display]:::secondary
G[Similarity Heatmap]:::secondary
H[Export Options]:::action
I[Download]:::action
J[Share]:::action
K[API Documentation]:::api
M[Integration Hub]:::api
subgraph P2P[P2P Controls]
direction TB
N[Protocol Selector]:::p2p
O[Peer Manager]:::p2p
P[Network Stats]:::p2p
Q[Storage Config]:::p2p
end
end
subgraph VCS[Version Controls]
direction TB
V[Branch Manager]:::version
W[Merge UI]:::version
X[History Browser]:::version
Y[Diff Viewer]:::version
L[Version History]:::version
V --> W
V --> X
X --> Y
X --> L
end
A --> B
A --> C
A --> K
A --> M
A --> P2P
A -.-> VCS
B --> D
C --> E
C --> F
C --> G
E & F & G --> H
H --> I
H --> J
N --> O
O --> P
N --> Q
M --> |External Systems| K
O --> |Network Status| P
style UI fill:#1a1c1d,stroke:#42b883,stroke-width:2px
style P2P fill:#1a1c1d,stroke:#00b894,stroke-width:2px
style VCS fill:#1a1c1d,stroke:#fd79a8,stroke-width:2px
%%{init: {'theme': 'dark', 'themeVariables': { 'darkMode': true }}}%%
flowchart TD
classDef engine fill:#2b7489,stroke:#1a1c1d,stroke-width:2px,color:white
classDef ai fill:#6b9fff,stroke:#2d5a9e,stroke-width:2px,color:#1a1c1d
classDef storage fill:#ff6b6b,stroke:#c92a2a,stroke-width:2px,color:white
classDef dist fill:#a55eea,stroke:#8854d0,stroke-width:2px,color:white
classDef version fill:#20bf6b,stroke:#26de81,stroke-width:2px,color:#1a1c1d
classDef p2p fill:#00b894,stroke:#00cec9,stroke-width:2px,color:#1a1c1d
classDef adapter fill:#6c5ce7,stroke:#a29bfe,stroke-width:2px,color:white
subgraph Core[Deduplication Core]
direction TB
A[Smart Classifier]:::engine
B[Deduplication Engine]:::engine
C[AI Processing Pipeline]:::ai
D1[ML Text Deduper]:::ai
D2[GNN JSON Deduper]:::ai
D3[CNN Image Deduper]:::ai
D4[Binary Deduper]:::ai
E[ML Conflict Resolver]:::ai
F[Auto-Tuning]:::ai
G[Vector Indexing]:::engine
H1[FAISS HNSW]:::engine
H2[Perceptual Hash]:::engine
H3[Graph Matching]:::engine
I[Storage Layer]:::storage
J1[Sled DB]:::storage
J2[File System]:::storage
J3[Vector DB]:::storage
%% Core Components
K[Model Registry]:::ai
L[Feedback Loop]:::ai
M[API Gateway]:::engine
%% P2P Components
Q[P2P Manager]:::p2p
R[Protocol Adapter]:::adapter
S1[LibP2P Implementation]:::p2p
S2[Holepunch Implementation]:::p2p
T[Peer Discovery]:::p2p
U[Content Router]:::p2p
V[Replication Manager]:::p2p
W[Storage Adapter]:::adapter
X1[IPFS Storage]:::p2p
X2[Hypercore Storage]:::p2p
A --> B
B --> C
C --> D1 & D2 & D3 & D4
D1 & D2 & D3 & D4 --> E
E --> F
B --> G
G --> H1 & H2 & H3
H1 & H2 & H3 --> I
I --> J1 & J2 & J3
%% Core Connections
E --> L
L --> K
K --> C
B --> M
%% P2P Connections
B <--> Q
Q --> R
R --> S1 & S2
S1 & S2 --> T
T --> U
U --> V
V --> W
W --> X1 & X2
X1 & X2 --> I
%% Feedback Loops
L --> |Model Updates| D1 & D2 & D3 & D4
F --> |Threshold Updates| L
M --> |API Requests| B
V --> |P2P Updates| L
end
subgraph VCS[Version Control Service]
direction TB
CRDT[Loro CRDT]:::version
BRANCH[Branch Manager]:::version
MERGE[Merge Handler]:::version
DAG[History DAG]:::version
BLOCK[Block Storage]:::version
CRDT --> BRANCH
CRDT --> MERGE
BRANCH --> DAG
MERGE --> DAG
DAG --> BLOCK
end
style Core fill:#1a1c1d,stroke:#2b7489,stroke-width:2px
style VCS fill:#1a1c1d,stroke:#20bf6b,stroke-width:2px
- Smart Classification: AI-powered detection and routing of different data types
- Multi-Modal Deduplication: Specialized engines for text, JSON, images, and binary data
- ML-Enhanced Matching: Combines exact and fuzzy matching using advanced AI models
- Intelligent Conflict Resolution: Machine learning-assisted conflict handling with human oversight
- High-Performance Search: Vector-based similarity search using FAISS HNSW
- Flexible Storage: Pluggable storage backends supporting Sled DB, file system, and vector databases
- P2P Capabilities: Decentralized data sharing and deduplication using IPFS and LibP2P
- CRDT-Based Version Control: Loro-powered versioning with automatic conflict resolution
- Rich Data Types: Support for complex data structures with semantic merging
- Modern, responsive UI for data visualization and management
- AI confidence score displays and similarity heatmaps
- Interactive conflict resolution interface
- Export and download capabilities
- P2P network controls and protocol selection
-
Deduplication Engine:
- ML Text Deduper
- GNN JSON Deduper
- CNN Image Deduper
- Deep Learning Binary Deduper
-
Smart Indexing:
- FAISS HNSW Vector Search
- Perceptual Hashing
- Graph Matching Algorithms
-
ML Conflict Resolution:
- Confidence-weighted decision making
- Automated threshold tuning
- Human-in-the-loop review system
-
P2P Network Layer:
- Pluggable protocol architecture
- LibP2P implementation with IPFS storage
- Holepunch implementation with Hypercore storage
- Protocol-agnostic adapter interface
- Unified peer discovery and routing
- Cross-protocol replication support
-
Version Control:
- Loro CRDT (v1.0+)
- Rich CRDT types
- High-performance DAG
- Semantic merging
- Block-level storage
- Sled DB for high-performance local storage
- File system integration for simple deployments
- Vector DB support (Weaviate/Pinecone)
- IPFS for LibP2P-based storage
- Hypercore for Holepunch-based storage
- Loro block storage for versions
[Coming Soon]
[Coming Soon]
[Coming Soon]
[Coming Soon]
[Coming Soon]
[Coming Soon]