Polygres

Polygres is an all-in-one database platform built by Evokoa for the Agent Era. It combines PostgreSQL hosting with native graph traversal, HNSW vector search, and hybrid retrieval APIs, aiming to support structured records, connected relationships, and semantic meaning in one system.

The product includes pgGraph and pgVector components. pgGraph is used to model relationships natively, traverse connections, discover neighbors, and map entity graphs without recursive SQL joins. pgVector stores and searches embeddings directly in the database, using HNSW indexing for similarity lookups and scalar filtering to narrow results. Polygres also describes hybrid search that brings Postgres, pgVector, and pgGraph together, and its demos show joint search across semantic vector spaces, relational metadata structures, and graph relationships in a single operation.

Polygres is presented for applications and AI agents. The site highlights a user-memories retrieval playground and says the CLI is built for speed, automation, and AI agents. Developers can manage projects, import data, and run queries from the terminal with Polygres CLI, use a visual console and schema playground to configure relationships, manage joint tables, and build hybrid indexes, or integrate through the Evokoa/polygres-sdk open-source SDK. The platform offers interactive demos, including a Wikipedia-scale hybrid index demo and a user-memory demo.

Deployment options include self-hosting with pgGraph and pgVector or using Polygres’s managed cloud service. The cloud is described as scalable and built with native Kubernetes and Docker scaling support, and the product includes a full PostgreSQL 17 instance for use with existing ORMs, clients, and tools. All core components are open source. Pricing and account details are linked from the site, but no specific plan terms appear in the provided text.