What is Polygres?

Polygres is a platform that transforms your existing PostgreSQL database into an internal search engine for AI agents. It combines graph traversal and multiple search methods (dense vectors, sparse/full-text, discovery/lookup) into a single hybrid query, returning ranked, token-ready context blocks. The underlying engines, pgGraph and pgContext, are open-source Postgres extensions.

How to use Polygres?

Start by hosting your database on Polygres or connecting an existing Postgres instance (e.g., Supabase, Neon) via a connection string. Install the Polygres CLI (pip install polygres-cli) to manage projects, import data, and run queries from the terminal. Use the SDK to call polygres.retrieve() with your query, graph hops, filters, and token limits to get ranked context for your AI agent.

Core features of Polygres

Hybrid search: combines graph traversal, dense vector, sparse/full-text, and discovery/lookup in one query
Graph engine (pgGraph): compiles foreign keys into a graph for multi-hop traversal, shortest paths, and neighborhood search
AI-native search (pgContext): ten search methods with fusion re-scoring
Token-aware context assembly: returns ranked, token-ready context blocks
Self-hosted or managed cloud: open-source extensions for any Postgres instance
CLI and SDK for easy integration