Evokoa/pgGraph
Open-source graph database superpowers for your existing Postgres data.
What it solves
Relational databases like PostgreSQL are excellent for structured data but struggle with complex graph-style questions—such as finding the shortest path between two entities or discovering records within a specific number of hops. Traditionally, these queries require complex, custom recursive SQL that can be slow and difficult to maintain.
How it works
Instead of moving data to a separate graph database, pgGraph builds a specialized, read-heavy graph index derived from your existing PostgreSQL tables. It uses a Compressed Sparse Row (CSR) format to store relationships in a contiguous memory structure, allowing for extremely fast traversals. You interact with this graph using standard SQL functions within the graph schema, while PostgreSQL remains the authoritative source of truth for your data.
Who it’s for
- Developers using PostgreSQL who need to perform complex relationship queries without migrating to a dedicated graph database.
- Data engineers looking to accelerate graph-based workloads like GraphRAG.
- Users who want to perform high-speed traversals (like shortest path or multi-hop searches) while keeping all data within their existing relational schema and security constraints.
Highlights
- Zero Data Migration: Your data stays in standard PostgreSQL tables; pgGraph only creates a derived index.
- High Performance: Uses a specialized memory-efficient execution layer to bypass the overhead of recursive SQL joins.
- SQL-Native: No new query language required; use SQL functions to perform graph searches.
- Safety Features: Includes built-in circuit breakers like depth limits and memory safeguards to prevent unbounded traversals from crashing the database.