16 Data Paradigms, 15 Live Today. One Connection. One Theoretical Framework. SQL, Document, Vector, Graph, and more--unified under one connection. Native PostgreSQL ACID for SQL operations; the Transaction paradigm (cross-model atomic workflows) is coming soon.
Fundamental paradigms, unified under one framework.
PolyDB doesn't just combine databases; it unifies the underlying data theory. Each model is optimized for its domain while sharing the same backend, security layers, and connection management.
Intelligent defaults out of the box. Start coding instantly with any model.
from polydb import PolyDB, PolyDBConfig # Unified entry point (OSS package, runs locally) db = PolyDB(PolyDBConfig()) # One connection, every model — lazy provider accessors db.sql().execute_query("SELECT 1") db.document().insert_document("profiles", {"name": "Ada"}) db.vector().search("embeddings", [0.1, 0.2, 0.7], k=5) db.graph().create_node("n1", "person", {"name": "Ada"}) # On the cloud, reach the same models over MCP today (hosted SDK soon).
Drop in one line. Works with Claude Code, Cursor, Codex, and any MCP-aware client.
PolyDB ships an MCP server exposing 15 of the 16 data providers as tools today (Transaction coming soon). Connect your AI coding assistant to a dedicated multi-model database in seconds -- no SDK, no boilerplate, no sprawl.
claude mcp add polydb \ --transport http \ https://beta.polydb.dev/mcp \ --header "X-PolyDB-Token: polydb_mcp_<your-token-here>"
Or any MCP client
Simple. Transparent. No surprises.
Enterprise / SLA? Contact us ->