🗄️ ThemisDB

High-Performance Multi-Model Database with Native AI/LLM Integration

"ThemisDB keeps its own llamas." – Run LLaMA, Mistral, Phi-3 directly in your database, no API calls needed.

---

🎉 What's New

🚀 v1.4.0-alpha - Advanced LLM Features (2026-01-05)

🧠 Advanced LLM Capabilities

• 📝 Grammar-Constrained Generation - EBNF/GBNF support for guaranteed valid JSON/XML/CSV outputs (95-99% reliability vs 60-70%)
  • Built-in grammars: JSON, XML, CSV, ReAct Agent
  • Thread-safe grammar cache with LRU eviction
  • Zero post-processing required
• 🔭 RoPE Scaling - Extended context window from 4K → 32K tokens (8x increase)
  • Linear, NTK-aware, YaRN scaling methods
  • Process entire research papers and codebases
• 🖼️ Vision Support - Multi-modal LLMs with CLIP-based image encoding
  • LLaVA integration for image analysis
  • Single and multiple image support
  • Thread-safe VisionEncoder class
• ⚡ Flash Attention - CUDA kernels for 15-25% speedup, 30% memory reduction
  • Optimized attention mechanism
  • Backward pass for training support
  • Multi-compute capability support
• 🎯 Speculative Decoding - 2-3x faster inference with draft+target models
• 🔄 Continuous Batching - 2x+ throughput with dynamic request batching

🏢 Enterprise Enhancements

• 🔥 Hot Spare Management - Automatic failover with health monitoring
• 📊 Enhanced Prometheus Metrics - LLM inference, cache performance, response tracking
• 🔄 WAL Replication via gRPC - Distributed inter-shard replication
• 🎮 Multi-GPU LoRA Support - Distributed LoRA adapters across GPUs
• 🐘 PostgreSQL Protocol - COPY, prepared statements, transaction support

📊 Quality & Testing

• 31 new test suites with comprehensive coverage
• 11 new performance benchmarks
• 17 new documentation guides
• 938 files changed (+113,762 lines, -45,154 lines)

📚 Quick Links:

• Grammar-Constrained Generation
• RoPE Scaling Guide
• Vision Support Quick Start
• Flash Attention Implementation
• Complete Changelog

---

�️ RAID Sharding Deadlock Hotfix (2026-01-04)

• 🔧 Critical Fix - Resolved server hang in RAID cluster mode at "Adaptive Index Manager initialized"
• 🎯 Root Cause - AdaptiveIndexManager MVCC coordination across 2 CFs before Sharding Manager initialization
• ✅ Solution - Conditional Column Family opening when THEMIS_ENABLE_SHARDING=true detected
• 🔧 Port Mapping Fix - Corrected docker-compose HTTP mappings from 808X:8080 to 808X:8765
• 🧪 RAID Testing - 2-hour endurance test suite with monitoring dashboard
• 📊 Verification - All 9 RAID shards (RAID 0/1/5) operational with 0% error rate
• 📦 Build Optimization - Docker build context reduced from 3GB to 85MB (97% reduction)

Files Modified: src/storage/rocksdb_wrapper.cpp, src/server/http_server.cpp, docker/compose/docker-compose-sharding.yml, Dockerfile.themis-server, .dockerignore
Tools Added: scripts/raid_endurance_test.py, scripts/monitor_raid_test.ps1

🔒 Security Improvements Summary (v1.3.0 - v1.3.4)

• 🛡️ Critical Security Fixes - Addressed 7 critical and 8 medium severity issues in RocksDB wrapper (100% segfault risk elimination)
• 🐳 Docker Security - Upgraded to Ubuntu 24.04 LTS with 80%+ CVE reduction
• 🔄 Update Checker Security - Secure token handling, HTTPS-only, thread-safe implementation
• 🔐 Binary Authenticity - Cryptographic manifest signing architecture (RSA-4096, SHA-256)
• 🔧 Memory Safety - Fixed use-after-free vulnerabilities, memory leaks, and resource management
• ✅ Verification - CodeQL passed, comprehensive audit reports, all critical issues resolved

Quick Links: Security Summary (EN) | Sicherheitszusammenfassung (DE) | Security Audit | Security Policy

🎙️ Voice Assistant Integration (2025-12-30) - Enterprise Feature

• 🗣️ Natural Language Voice Interaction - Similar to Alexa/Siri, powered by Whisper.cpp + Piper TTS + llama.cpp
• 📞 Phone Call Recording - Automatic transcription and secure storage with revision control
• 📝 Meeting Protocol Generation - AI-powered meeting minutes, key points, and action items
• 🎯 Speaker Diarization - Identify and label different speakers in recordings
• 🔒 Revision-Safe Storage - All recordings stored securely in ThemisDB with full audit trails
• 🌐 Multi-Language Support - 100+ languages for STT, multiple voices for TTS

Quick Links: Voice Assistant Guide (EN) | Sprachassistent (DE)

🌿 Git Flow Branching Strategy (2025-12-30)

• 🎯 Git Flow Implementation - main as production release branch, develop as integration branch
• 📚 Comprehensive Documentation - Bilingual guides (DE/EN), visual workflows, quick reference cards
• 🚀 Branch-Based CI/CD - Fast builds on develop (~5-10 min), production builds on main (~30-40 min)
• 🔄 Migration Guide - Step-by-step transition for existing contributors
• 🛡️ Branch Protection Setup - GitHub configuration with automated scripts

Quick Links: Branching Strategy (DE) | Branching Strategy (EN) | Quick Reference | Documentation Hub

v1.3.3 - Network Protocol Enhancements (2025-12-21)

• 🌐 HTTP/2 with Server Push for CDC/Changefeed with ~0ms latency
• 📡 WebSocket support with CDC streaming for real-time communication
• 📬 MQTT broker with WebSocket transport and monitoring
• ⚡ HTTP/3 base implementation with QUIC protocol
• 🐘 PostgreSQL Wire Protocol for BI tool compatibility
• 🔌 MCP Server (Model Context Protocol) support

v1.3.2 - Image Analysis AI Plugin (2025-12-21)

• 🖼️ Image Analysis AI Plugin Architecture running parallel with LLM
• 🔧 Multi-backend support: llama.cpp Vision, ONNX Runtime, OpenCV DNN, OpenVINO, ncnn
• 🛠️ Plugin interfaces: IImageAnalysisBackend, ImageAnalysisManager
• 🧪 Comprehensive unit tests (15+ test cases) and benchmarks

v1.3.1 - Third-Party Attribution (2025-12-20)

• 📄 Added ATTRIBUTIONS.md documenting 15+ core dependencies
• 🏆 Documented ThemisDB's 12 unique innovations
• ✅ Clear attribution for all major dependencies

v1.3.0 - LLM Integration (2025-12-17)

🧠 Native LLM Integration with llama.cpp (Optional)

"ThemisDB keeps its own llamas." – Run AI/LLM workloads directly in your database - no external API costs!

[!NOTE] LLM integration is an optional feature that requires: -# LLM Features (When Enabled)

Feature	Description	Status
🧠 Embedded LLM Engine	llama.cpp integration for LLaMA/Mistral/Phi-3 (1B-70B params)	✅
📝 Grammar Constraints	EBNF/GBNF for guaranteed valid JSON/XML/CSV outputs	✅ v1.4.0-alpha
🔭 RoPE Scaling	Extended context window 4K → 32K tokens (8x increase)	✅ v1.4.0-alpha
🖼️ Vision Support	Multi-modal LLMs with CLIP image encoding (LLaVA)	✅ v1.4.0-alpha
⚡ Flash Attention	CUDA kernels: 15-25% speedup, 30% memory reduction	✅ v1.4.0-alpha
🎯 Speculative Decoding	2-3x faster inference with draft+target models	✅ v1.4.0-alpha
🔄 Continuous Batching	2x+ throughput with dynamic request batching	✅ v1.4.0-alpha
🎙️ Voice Assistant	STT/TTS/LLM for phone calls, meetings, voice commands (Enterprise)	✅
🖼️ Image Analysis AI	Multi-backend plugins (llama.cpp Vision, ONNX CLIP, OpenCV DNN)	✅
⚡ GPU Acceleration	NVIDIA CUDA support with significant speedup	✅
💾 PagedAttention	Advanced memory management	✅
🔧 Quantization	Q4_K_M, Q5_K_M, Q8_0 for efficient memory usage	✅
📊 Monitoring	Grafana dashboards with metrics and alerts	✅
🔌 Plugin Architecture	Extensible LLM and image analysis backends	✅
🌐 Distributed RPC	Inter-shard communication for distributed LLM ops	✅

Performance Highlights

[!TIP] GPU acceleration provides significant speedup over CPU with PagedAttention memory savings.

• ⚡ Significant speedup with GPU acceleration vs CPU
• 💾 Memory savings with PagedAttention and prefix caching
• 🚀 Kernel fusion for additional performance gains
• ✅ Comprehensive test coverage with unit tests

📚 Documentation:

• 🧠 LLM Complete Setup Guide (DE) - Vollständiger Guide für LLM-Setup und Inferencing
• 🎯 Overview - System architecture and design

ThemisDB is a production-ready multi-model database that combines relational, graph, vector, and document models in a single system with full ACID transaction support. Built on RocksDB with advanced security and compliance features.

📦 Editions

Available Editions

Edition	License	Features	Use Case
🔹 Minimal	Open Source (MIT)	Core database only - no LLM, GPU, sharding, advanced protocols	Embedded systems, IoT, edge devices, fast builds
🆓 Community	Open Source (MIT)	Full-featured single-node database with all core capabilities	Development, startups, single-server deployments
🔒 Enterprise	Commercial	+ Horizontal scaling, advanced analytics, HA/replication, and more	Large-scale production deployments

→ See Minimal Edition Details | → See Enterprise Edition Details

---

✨ Features

🔑 Core Features

Database Capabilities

Feature	Description	Community	Enterprise
🚀 Quick Start

🐳 Docker (Recommended)

# Pull and run the latest version
docker pull themisdb/themisdb:latest

# Run with Docker
docker run -d \
  --name themis \
  -p 8080:8080 \
  -p 18765:18765 \
  -p 4318:4318 \
  -v themis_data:/data \
  themisdb/themisdb:latest

# Or use Docker Compose
docker compose up -d

# Verify installation
curl http://localhost:8080/health

[!TIP] Use Docker Compose for production deployments with proper configuration.

📡 Default Ports

Port	Protocol	Description
8080	HTTP/1.1	REST API, GraphQL
18765	Binary	Wire Protocol, gRPC
4318	HTTP	OpenTelemetry/Prometheus

[!NOTE] Complete Port Reference: See [v1.3.0+)

• ✅ Image Analysis - Multi-backend AI plugins (v1.3.0+)
• ✅ GNN Embeddings - Graph Neural Network support

🌐 Modern Protocols

Protocol	Status	Description
HTTP/1.1	✅	REST API, GraphQL
HTTP/2	✅	Server Push for CDC
HTTP/3	🚧	QUIC (experimental)
WebSocket	✅	Bidirectional streaming
gRPC	✅	Binary RPC
MQTT	✅	IoT messaging
PostgreSQL Wire	✅	BI tool compatibility
MCP	✅	Model Context Protocol
SSE	✅	Server-Sent Events

📚 Transparency & Attribution

ThemisDB is built on proven open-source foundations with clear attribution:

• ✅ Transparent Attribution - Clear documentation of all dependencies
• ✅ Innovation Documentation - ThemisDB's unique contributions vs third-party features
• ✅ License Compliance - Full license information for all components

→ See Complete Attribution

**Key Features:**

• 🔒 ACID Transactions - Full snapshot isolation with MVCC
• 🔍 Multi-Model - Relational, Graph, Vector, Document in one database
• 🚀 High Performance - 45K writes/s, 120K reads/s, GPU-accelerated vector search
• 🛡️ Security - TLS 1.3, RBAC, field-level encryption, audit logging (Enterprise: HSM integration)
• 📊 Analytics - Time-series, aggregations (Enterprise: OLAP, CEP, materialized views)
• 🌐 Distribution - Single-node optimized (Enterprise: horizontal sharding, replication, Kubernetes)
• 🧠 AI-Ready - Hybrid search (RAG), embedding cache, FAISS integration, optional LLM engine with llama.cpp (v1.3.0+), image analysis AI plugins (v1.3.0+), voice assistant with STT/TTS (v1.4.0+, Enterprise)
• 🌐 Modern Protocols - HTTP/1.1, GraphQL, SSE, gRPC (v1.3.0), HTTP/2 with Server Push ✅, WebSocket ✅, MQTT ✅, HTTP/3 🚧, PostgreSQL Wire ✅, MCP ✅
• 📚 Transparent Attribution - Clear documentation of third-party dependencies vs ThemisDB innovations (see ATTRIBUTIONS.md)
• 🖼️ Image Analysis - Multi-backend AI plugin architecture (llama.cpp Vision, ONNX CLIP, OpenCV DNN)

---

Quick Start

Docker (Recommended)

# Pull and run the latest stable version
docker pull themisdb/themisdb:latest
docker run -d \
  -p 8080:8080 \
  -p 18765:18765 \
  -p 4318:4318 \
  -v themis_data:/data \
  themisdb/themisdb:latest

# Or use nightly builds (latest development version)
docker pull themisdb/server:nightly
docker run -d -p 8080:8080 -p 18765:18765 -v themis_data:/data themisdb/server:nightly

# Or use Docker Compose
docker compose up -d

# Check health
curl http://localhost:8080/health

Default Ports:

• 8080 - HTTP/REST API, GraphQL
• 18765 - Binary Wire Protocol, gRPC
• 4318 - OpenTelemetry/Prometheus metrics

📖 Complete Port Reference: See docs/deployment/PORT_REFERENCE.md for all ports including optional protocols (MQTT, PostgreSQL Wire, MCP).

Release Builds: Automated release builds with edition support (community/enterprise/hyperscaler) are available on DockerHub. Each release includes automatically generated changelogs and enterprise security artifacts (SBOM, signatures, SLSA provenance). See Release Build Documentation and Release Changelogs for details.

From Source

# Clone repository
git clone https://github.com/makr-code/ThemisDB.git
cd ThemisDB

# Setup and build (Linux/macOS)
./scripts/setup.sh
./scripts/build.sh

# Setup and build (Windows)
.\scripts\setup.ps1
.\scripts\build.ps1

# Start server
./build/themis_server --config config.yaml

[!TIP] Configuration: See config/config.yaml for complete configuration reference with all available options, or check the Configuration Guide | Tuning Guide (DE) for detailed tuning recommendations.

Optional Protocol Support (Security: Opt-In by Default):

# Enable HTTP/2 with Server Push (explicit opt-in for security)
cmake -B build -S . -DTHEMIS_ENABLE_HTTP2=ON

# Enable WebSocket with CDC (explicit opt-in for security)
cmake -B build -S . -DTHEMIS_ENABLE_WEBSOCKET=ON

# Enable MQTT broker (explicit opt-in for security)
cmake -B build -S . -DTHEMIS_ENABLE_MQTT=ON

# Enable PostgreSQL Wire Protocol (explicit opt-in for security)
cmake -B build -S . -DTHEMIS_ENABLE_POSTGRES_WIRE=ON

# Enable MCP for LLM integration (explicit opt-in for security)
cmake -B build -S . -DTHEMIS_ENABLE_MCP=ON

# Enable HTTP/3 (explicit opt-in for security)
cmake -B build -S . -DTHEMIS_ENABLE_HTTP3=ON

# Default build only includes HTTP/1.1, GraphQL, SSE, gRPC (minimal attack surface)

See Protocol Documentation for details.

Windows: Build mit LLM (llama.cpp) - Optional

# OPTIONAL: Für LLM-Unterstützung - lokaler Clone von llama.cpp erforderlich
if (!(Test-Path "llama.cpp")) {
  git clone https://github.com/ggerganov/llama.cpp.git llama.cpp
}

# MSVC Release-Build mit LLM-Unterstützung
powershell -File scripts/build-themis-server-llm.ps1

# Sanity-Check
./build-msvc/bin/themis_server.exe --help

Hinweise:

• LLM-Unterstützung ist optional und erfordert -DTHEMIS_ENABLE_LLM=ON beim Build
• llama.cpp/ liegt als lokaler Clone im Projekt-Root und ist per .gitignore und .dockerignore ausgeschlossen (wird nicht committed oder in Docker kopiert)
• Der Build-Skript setzt Visual Studio 2022 (-G "Visual Studio 17 2022") und -A x64, bindet die vcpkg-Toolchain ein und behebt MSVC‑spezifische char8_t‑Fehler am llama‑Target

→ Comprehensive Build Documentation | Build-Varianten, Plattformen, Troubleshooting

Package Managers

Linux (Debian/Ubuntu):

wget https://github.com/makr-code/ThemisDB/releases/latest/download/themisdb_1.3.0-1_amd64.deb
sudo apt install ./themisdb_1.3.0-1_amd64.deb
sudo systemctl start themisdb

macOS (Homebrew):

brew install themisdb
brew services start themisdb

Windows (Chocolatey):

choco install themisdb

---

5-Minute Tutorial

# 1. Check server health
curl http://localhost:8765/health

# 2. Create an entity
curl -X PUT http://localhost:8765/entities/users:alice \
  -H "Content-Type: application/json" \
  -d '{"blob":"{\"name\":\"Alice\",\"age\":30,\"city\":\"Berlin\"}"}'

# 3. Create an index
curl -X POST http://localhost:8765/index/create \
  -H "Content-Type: application/json" \
  -d '{"table":"users","column":"city"}'

# 4. Query by index
curl -X POST http://localhost:8765/query \
  -H "Content-Type: application/json" \
  -d '{"table":"users","predicates":[{"column":"city","value":"Berlin"}],"return":"entities"}'

# 5. View metrics
curl http://localhost:8765/metrics

---

Architecture

ThemisDB uses a unified storage architecture with specialized projection layers:

┌─────────────────────────────────────────────────────────┐
│                   Query Layer (AQL)                     │
│  SQL-like • Graph Traversals • Vector Search • Analytics│
├─────────────────────────────────────────────────────────┤
│                 Projection Layers                        │
│  Secondary Indices • Graph Adjacency • HNSW Vector      │
├─────────────────────────────────────────────────────────┤
│              Canonical Storage (Base Entity)             │
│         RocksDB LSM-Tree • MVCC Transactions            │
└─────────────────────────────────────────────────────────┘

Core Components:

• Storage Engine: RocksDB TransactionDB with LSM-Tree
• Transaction Manager: MVCC with snapshot isolation
• Query Engine: Advanced Query Language (AQL) with graph/vector support
• Index Manager: Automatic maintenance of secondary, graph, and vector indexes
• Security: TLS 1.3, RBAC, field encryption, audit logging
• Observability: Prometheus metrics, OpenTelemetry tracing

→ Full Architecture Documentation

---

Core Features

Multi-Model Database

• Relational: SQL-like queries with secondary indexes
• Graph: BFS, Dijkstra, A* traversals with path constraints
• Vector: HNSW and FAISS for similarity search (GPU-accelerated)
• Document: JSON storage with flexible schema
• Time-Series: Gorilla compression, continuous aggregates

Transaction Support

• Full ACID guarantees with snapshot isolation
• Write-write conflict detection
• Atomic updates across all index types
• Session-based and direct API

Advanced Analytics

• CEP Engine: Complex Event Processing with pattern matching
• OLAP: CUBE, ROLLUP, window functions
• Time-Series: Compression, retention policies, aggregates
• Hybrid Search: BM25 + vector for RAG workflows

Enterprise Security

• TLS 1.3 with mTLS support
• Role-Based Access Control (RBAC)
• Field-level encryption
• Audit logging with SIEM integration
• Certificate pinning for HSM/TSA
• Secrets management (HashiCorp Vault)

Distributed Capabilities

• Horizontal sharding with consistent hashing
• Leader-follower and multi-master replication
• RAID-like redundancy (MIRROR, STRIPE, PARITY)
• Kubernetes operator with CRDs
• Auto-rebalancing and cloud deployment

GPU Acceleration (Optional)

• 10 backend options: CUDA, Vulkan, HIP, OpenCL, DirectX, OneAPI, ZLUDA
• 10-50x speedup for vector search
• Automatic platform detection and fallback

---

Documentation

Getting Started:

Core Concepts:

Features:

Operations:

Development:

Full Documentation: https://makr-code.github.io/ThemisDB/

📚 Documentation

Getting Started

• 🚀 Quick Start
• 📖 5-Minute Tutorial
• 🐳 Docker Deployment
• 🔧 Building from Source

Core Concepts

• 🏗️ Architecture Overview
• 💾 Multi-Model Design
• 🔄 Transaction Management
• 🔍 AQL Query Language

Features

• 🎯 Vector Search
• 🕸️ Graph Operations
• 📈 Time-Series Engine
• 🔐 Security & Compliance
• ⚡ Feature Overview

Operations

• ⚙️ Configuration Guide | Tuning Guide (DE)
• 📊 Monitoring & Metrics
• 💾 Backup & Recovery
• ⚡ Performance Tuning

Development

• 🔨 Build Guide
• 🌿 Branching Strategy | EN
• 🤝 Contributing
• 📖 API Reference
• 📦 Client SDKs

Enterprise & Strategy

• 🏛️ CMS Strategy Paper (DE) - ThemisDB für Content Management in Government und Enterprise
• 💼 Enterprise Edition - Enterprise features and licensing
• 📊 Governance - Data governance and policies

[!NOTE] Full Documentation: https://makr-code.github.io/ThemisDB/

---

🗺️ Roadmap

✅ Completed (v1.0 - v1.3)

Production-Ready Features

• ✅ ACID transactions with MVCC
• ✅ Multi-model support (relational, graph, vector, document)
• ✅ Horizontal sharding and replication
• ✅ GPU acceleration (10 backends)
• ✅ Enterprise security features
• ✅ Client SDKs (7 languages)
• ✅ Kubernetes operator
• ✅ Native LLM integration (optional)
• ✅ Modern protocol support (HTTP/2, WebSocket, gRPC, MQTT, PostgreSQL Wire, MCP)

🚧 In Progress (v1.4 - Q1 2026)

• 🚧 Query Optimizer - Advanced query optimization and execution plans
• 🚧 Multi-Datacenter - Cross-region deployment support
• 🚧 Advanced ML/GNN - Enhanced machine learning features
• 🚧 Production Hardening - Additional stability and performance improvements

📋 Planned (v1.5+ - 2026)

• 📋 Modular Architecture - Split monolithic core into 11 focused libraries
• 📋 Real-Time Views - Materialized views with automatic updates
• 📋 Cross-Region Replication - Global data distribution
• 📋 Advanced Compliance - SOC 2, HIPAA certification
• 📋 Cloud-Native Optimizations - Enhanced cloud provider integrations

📚 Detailed Planning:

• Complete Roadmap
• Modularization Plan

---

⚡ Performance

Benchmark Results

Test Environment: Release build, Windows x64, 20 cores @ 3696 MHz

Operation	Throughput	Latency (avg)	Notes
📝 Entity PUT	45,000 ops/s	0.02 ms	Write throughput
📖 Entity GET	120,000 ops/s	0.008 ms	Read throughput
🔍 Indexed Query	3.4M queries/s	0.29 μs	AQL WHERE clause
🕸️ Graph Traverse	9.56M ops/s	0.105 μs	BFS (depth=3)
🎯 Vector Search (RGB)	59.7M queries/s	0.017 μs	Simple 3D vectors
📊 Vector Insert (384D)	411k vectors/s	2.44 μs	Typical embeddings
🧠 RAG Search (Top-50)	7.17M queries/s	0.14 μs	LLM retrieval

[!IMPORTANT] Performance Disclaimer: Benchmarks represent optimal conditions. Actual performance varies based on:

• Hardware configuration (CPU, RAM, storage)
• Data size and complexity
• Concurrent workload patterns
• Build configuration and optimizations

📊 Detailed Analysis:

• Complete Benchmark Results
• [🤝 Community & Support

Resource	Description	Link
📚 Documentation	Complete guides and API reference	Docs Site
🐛 Issues	Report bugs or request features	GitHub Issues
💬 Discussions	Community Q&A and discussions	GitHub Discussions
🤝 Contributing	How to contribute to ThemisDB	Contributing Guide
🌿 Branching Strategy	Git Flow workflow (main = release, develop = dev)	Branching Strategy
🔒 Security	Responsible disclosure policy	Security Policy

---

📄 License

License Information

Community Edition

ThemisDB Community Edition is released under the MIT License.

• ✅ Free to use, modify, and distribute
• ✅ Commercial use allowed
• ✅ Full feature set for single-node deployments

Enterprise Edition

ThemisDB Enterprise Edition features (horizontal sharding, advanced analytics, HA/replication, etc.) are available under a commercial license.

Enterprise Inquiries: [email protected]

→ See Enterprise Features

---

🙏 Acknowledgments

ThemisDB builds upon and is inspired by these excellent projects:

Inspirations & Foundations

Project	Influence	Area
ArangoDB	Multi-model architecture	Design Philosophy
CozoDB	Hybrid relational-graph-vector	Data Models
Azure Cosmos DB	Multi-model with unified API	API Design
RocksDB	High-performance LSM-Tree storage	Storage Engine
FAISS	Efficient similarity search	Vector Search

[!NOTE] For a complete list of third-party libraries and detailed feature attributions, see ATTRIBUTIONS.md.

---

Built with ❤️ for the database community

⭐ Star us on GitHub · 📖 Read the Docs · 🤝 Contribute

es and feature attributions, see [ATTRIBUTIONS.md](ATTRIBUTIONS.md).**

---

Built with ❤️ for the database community