Cognee Community ArcadeDB Hybrid Adapter

A community-maintained adapter that enables Cognee to work with ArcadeDB as a unified hybrid database — both graph and vector operations in a single engine.

Based on PR#94 by @lvca (ArcadeDB team) with additional fixes for Cognee 1.0+ compatibility and production reliability.

Installation

pip install cognee-community-hybrid-adapter-arcadedb

# Optional: Bolt protocol support (faster graph operations)
pip install "cognee-community-hybrid-adapter-arcadedb[bolt]"

Usage

import asyncio
import os
import pathlib
from os import path
from cognee import config, prune, add, cognify, search, SearchType

# Import the register module to enable ArcadeDB support
from cognee_community_hybrid_adapter_arcadedb import register


async def main():
    # Set up local directories
    system_path = pathlib.Path(__file__).parent
    config.system_root_directory(path.join(system_path, ".cognee_system"))
    config.data_root_directory(path.join(system_path, ".cognee_data"))

    # Configure databases
    config.set_relational_db_config(
        {
            "db_provider": "sqlite",
        }
    )

    # Configure ArcadeDB as both vector and graph database
    config.set_vector_db_config(
        {
            "vector_db_provider": "arcadedb",
            "vector_db_url": os.getenv("ARCADEDB_URL", "localhost"),
            "vector_db_port": int(os.getenv("ARCADEDB_HTTP_PORT", "2480")),
            "vector_db_username": os.getenv("ARCADEDB_USERNAME", "root"),
            "vector_db_password": os.getenv("ARCADEDB_PASSWORD", ""),
        }
    )
    config.set_graph_db_config(
        {
            "graph_database_provider": "arcadedb",
            "graph_database_url": os.getenv("ARCADEDB_URL", "localhost"),
            "graph_database_port": int(os.getenv("ARCADEDB_HTTP_PORT", "2480")),
            "graph_database_username": os.getenv("ARCADEDB_USERNAME", "root"),
            "graph_database_password": os.getenv("ARCADEDB_PASSWORD", ""),
        }
    )

    # Optional: Clean previous data
    await prune.prune_data()
    await prune.prune_system(metadata=True)

    # Add and process your content
    await add("""
    Natural language processing (NLP) is an interdisciplinary
    subfield of computer science and information retrieval.
    """)

    await cognify()

    # Search using graph completion
    search_results = await search(
        query_type=SearchType.GRAPH_COMPLETION,
        query_text="Tell me about NLP",
    )

    for result in search_results:
        print("\nSearch result:\n" + result)


if __name__ == "__main__":
    asyncio.run(main())

Configuration

The ArcadeDB adapter can be configured as both a vector database and graph database, providing true hybrid capabilities.

Environment Variables

export ARCADEDB_URL="localhost"
export ARCADEDB_HTTP_PORT="2480"
export ARCADEDB_USERNAME="root"
export ARCADEDB_PASSWORD="your-password"
export GRAPH_DATASET_DATABASE_HANDLER="arcadedb_graph_local"
export VECTOR_DATASET_DATABASE_HANDLER="arcadedb_vector_local"

ArcadeDB Setup

ArcadeDB is a multi-model database that natively supports graph traversal, vector search, and document operations in a single engine.

HTTP only (simplest):

docker run -p 2480:2480 \
  -e JAVA_OPTS="-Darcadedb.server.rootPassword=your-password" \
  arcadedata/arcadedb:latest

With Bolt protocol (recommended for better graph performance):

docker run -p 2480:2480 -p 7687:7687 \
  -e JAVA_OPTS="-Darcadedb.server.rootPassword=your-password \
    -Darcadedb.server.plugins=Bolt:com.arcadedb.bolt.BoltProtocolPlugin" \
  arcadedata/arcadedb:latest

Configuration Parameters

Graph Database Configuration:

• graph_database_provider: Set to "arcadedb"
• graph_database_url: ArcadeDB server hostname (default: "localhost")
• graph_database_port: HTTP API port (default: 2480)
• graph_database_username: Database username
• graph_database_password: Database password

Vector Database Configuration:

• vector_db_provider: Set to "arcadedb"
• vector_db_url: ArcadeDB server hostname (default: "localhost")
• vector_db_port: HTTP API port (default: 2480)
• vector_db_username: Database username
• vector_db_password: Database password

How it works

Graph operations

Graph operations use the Neo4j Bolt protocol when the BoltProtocolPlugin is enabled (port 7687), with automatic fallback to the HTTP API (OpenCypher). The adapter verifies Bolt connectivity on first use and transparently falls back if unavailable.

Vector operations

Vector operations use ArcadeDB's HTTP API for SQL queries:

• Index creation: CREATE INDEX ... LSM_VECTOR METADATA { dimensions: N, similarity: 'COSINE' }
• Vector storage: Stored as ARRAY_OF_FLOATS properties on vertex records
• Search: SELECT expand(vectorNeighbors('Type[prop]', [vec], k)) returns results sorted by distance

Auto-detection

The adapter automatically:

• Detects the Cypher vertex type casing (Vertex vs vertex) based on ArcadeDB version
• Creates the target database on first use (if it doesn't exist)
• Retries on HTTP 503 ConcurrentModificationException (ArcadeDB uses optimistic concurrency)

Key differences from PR#94

This adapter is based on PR#94 by @lvca with the following improvements:

Requirements

• Python >= 3.11, <= 3.13
• ArcadeDB >= 26.3 (tested with 26.4.2)
• Cognee >= 1.0.3

About ArcadeDB

ArcadeDB is a multi-model database engine that combines the power of graph databases, document stores, and vector search in a single, unified system:

• Graph: Full OpenCypher and Gremlin support, with Neo4j Bolt protocol compatibility
• Vector: Native HNSW/LSM_VECTOR indexes with cosine/euclidean similarity
• Document: JSON document storage with SQL queries
• Multi-model: Use all paradigms together in a single query

Key benefits:

• No separate vector DB needed — keep embeddings co-located with the knowledge graph
• Distributed via Raft — built-in high availability (no external coordination service)
• Low latency — optimized for fast traversals and real-time analytics
• SQL + Cypher — query in either language, or mix them

Learn more at arcadedb.com and the GitHub repository.