VectorDB AI

Universal vector database adapter for AI applications

Switch between 9 vector database backends without changing your code. No vendor lock-in.

Features

✅ 9 Vector Database Backends

• Tier 1: ChromaDB, Pinecone, Qdrant (embedded + cloud)
• Tier 2: Elasticsearch, Solr (enterprise search)
• Tier 3: Weaviate, Milvus, pgvector, Redis (specialized)

✅ Portable Data Format

• Export from any backend
• Import to any backend
• JSON-based interchange

✅ Semantic Embeddings

• 384-dimensional vectors (all-MiniLM-L6-v2)
• Automatic embedding generation
• Configurable models

✅ Production Ready

• Abstract interface
• Type hints
• Comprehensive error handling

Quick Start

from vectordb_ai import VectorStoreFactory, ContentChunk, embed

# Create store (ChromaDB by default)
store = VectorStoreFactory.create(
    backend='chromadb',
    collection='my-docs'
)

# Add documents
chunks = [
    ContentChunk(
        chunk_id='doc1',
        content='Python is awesome',
        source='tutorial.pdf',
        embedding=embed('Python is awesome')
    )
]
store.add_chunks(chunks)

# Search
results = store.search_chunks(
    query_embedding=embed('programming languages'),
    top_k=5
)

Switch Backends

# Start with ChromaDB (embedded)
store = VectorStoreFactory.create(backend='chromadb', collection='docs')

# Export data
store.export('data.json')

# Switch to Pinecone (cloud)
store = VectorStoreFactory.create(
    backend='pinecone',
    collection='docs',
    api_key='your-key'
)
store.import_from('data.json')

Supported Backends

Backend	Type	Use Case	Install
ChromaDB	Embedded	Development, small-scale	pip install vectordb-ai
Pinecone	Cloud	Production, scalable	pip install vectordb-ai[pinecone]
Qdrant	Cloud/Self-hosted	Production	pip install vectordb-ai[qdrant]
Elasticsearch	Self-hosted	Enterprise	pip install vectordb-ai[elasticsearch]
Solr	Self-hosted	Enterprise	pip install vectordb-ai[solr]
Weaviate	Cloud/Self-hosted	GraphQL API	pip install vectordb-ai[weaviate]
Milvus	Self-hosted	High performance	pip install vectordb-ai[milvus]
pgvector	Postgres	Existing Postgres	pip install vectordb-ai[pgvector]
Redis	Self-hosted	In-memory	pip install vectordb-ai[redis]

Installation

# Basic (ChromaDB only)
pip install vectordb-ai

# With specific backend
pip install vectordb-ai[pinecone]
pip install vectordb-ai[qdrant]

# All backends
pip install vectordb-ai[all]

Part of FlossWare AI

• vectordb-ai - Vector database adapter (this project)
• semantic-search-ai - Search enhancements
• consensus-ai - Multi-AI orchestration
• knowledge-ai - Knowledge ingestion
• universal-ai - AI platform

License

GPL-3.0 - FlossWare (sfloess)