​
Weaviate Integration

Weaviate is an open-source vector database that supports multi-modal data and hybrid search capabilities. It combines vector search with traditional keyword search for comprehensive retrieval.
Weaviate excels at multi-modal applications where you need to search across text, images, and other data types with advanced filtering and GraphQL querying.

​
Why Weaviate?

Multi-Modal Support

Beyond Text
  • Text, image, and audio embeddings
  • Cross-modal search capabilities
  • Unified vector space
  • Rich media understanding

Hybrid Search

Best of Both Worlds
  • Vector similarity + keyword search
  • BM25 + semantic search
  • Configurable ranking algorithms
  • Advanced query capabilities

GraphQL API

Flexible Querying
  • Rich query language
  • Nested object support
  • Aggregations and analytics
  • Type-safe queries

Open Source

Full Control
  • Self-hosted deployment
  • Custom modifications
  • No vendor lock-in
  • Community support

​
Installation & Setup

1

Deploy Weaviate

Choose your deployment method:Docker (Development):
docker run -d \
  --name weaviate \
  -p 8080:8080 \
  -e QUERY_DEFAULTS_LIMIT=25 \
  -e AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED=true \
  -e PERSISTENCE_DATA_PATH='/var/lib/weaviate' \
  semitechnologies/weaviate:latest
Docker Compose:
version: '3.8'
services:
  weaviate:
    image: semitechnologies/weaviate:latest
    ports:
      - "8080:8080"
    environment:
      QUERY_DEFAULTS_LIMIT: 25
      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
      PERSISTENCE_DATA_PATH: '/var/lib/weaviate'
    volumes:
      - weaviate_data:/var/lib/weaviate
volumes:
  weaviate_data:
Weaviate Cloud: Sign up at console.weaviate.cloud
2

Install Dependencies

# Install Cognee with Weaviate support
pip install cognee
pip install weaviate-client
3

Configure Cognee

import os
import cognee

# Weaviate configuration
os.environ["VECTOR_DB_PROVIDER"] = "weaviate"
os.environ["VECTOR_DB_URL"] = "http://localhost:8080"

# For Weaviate Cloud
# os.environ["VECTOR_DB_URL"] = "https://your-cluster.weaviate.network"
# os.environ["WEAVIATE_API_KEY"] = "your-weaviate-api-key"

print("Weaviate configured successfully!")

​
Configuration Examples

import os
import cognee

# Local Weaviate instance
os.environ["VECTOR_DB_PROVIDER"] = "weaviate"
os.environ["VECTOR_DB_URL"] = "http://localhost:8080"

# Optional: Configure class name
os.environ["WEAVIATE_CLASS_NAME"] = "CogneeDocument"

# Test the setup
await cognee.add("Weaviate supports GraphQL queries and vector search.")
await cognee.cognify()

​
Multi-Modal Features

Cross-Modal Search
# Add text and image data
await cognee.add("A diagram showing neural network architecture.")
await cognee.add_image("path/to/neural_network_diagram.png")

# Cross-modal search
results = await cognee.search(
    "neural network visualization",
    query_type=SearchType.MULTI_MODAL,
    include_images=True
)

​
Performance Optimization

Index Tuning

​
Monitoring & Analytics

import cognee
import asyncio

async def get_weaviate_metrics():
    # Get comprehensive metrics
    metrics = await cognee.get_weaviate_metrics()
    
    print("=== Weaviate Performance Metrics ===")
    print(f"Total objects: {metrics['total_objects']:,}")
    print(f"Index size: {metrics['index_size_mb']:.1f} MB")
    print(f"Memory usage: {metrics['memory_usage_mb']:.1f} MB")
    print(f"Average query time: {metrics['avg_query_time_ms']:.1f} ms")
    print(f"Cache hit rate: {metrics['cache_hit_rate']:.1%}")
    
    print("\n=== Query Performance ===")
    print(f"Queries/second: {metrics['qps']:.1f}")
    print(f"P95 latency: {metrics['p95_latency_ms']:.1f} ms")
    print(f"Error rate: {metrics['error_rate']:.2%}")
    
    return metrics

asyncio.run(get_weaviate_metrics())

​
Troubleshooting

Cannot Connect
# Test Weaviate connection
import weaviate

try:
    client = weaviate.Client("http://localhost:8080")
    ready = client.is_ready()
    
    if ready:
        print("✅ Weaviate is ready")
        schema = client.schema.get()
        print(f"Schema classes: {len(schema['classes'])}")
    else:
        print("❌ Weaviate not ready")
        
except Exception as e:
    print(f"❌ Connection failed: {e}")

​
Next Steps

Vector Stores Overview

Compare OptionsCompare Weaviate with other vector store options for your use case.

Graph Stores

Graph IntegrationLearn how vector and graph stores work together in Cognee.

Multi-Modal Setup

Advanced FeaturesConfigure multi-modal search and cross-modal capabilities.

Production Deployment

Scale Your AppDeploy Weaviate + Cognee for production workloads.