EdgeDB (Gel) — The Next-Generation Graph-Relational Database

What it is

EdgeDB (recently rebranded as Gel) is an open-source database built on top of PostgreSQL. It replaces SQL with EdgeQL, a query language designed for deep nested fetches and graph traversals. Objects in EdgeDB have typed properties and first-class links to other objects, so relationships are part of the schema rather than join tables.

EdgeDB targets backend developers who are tired of ORM boilerplate and N+1 query problems. Because links are resolved at the query engine level, a single EdgeQL query can fetch deeply nested object graphs without manual joins.

How it saves time or tokens

EdgeDB eliminates the mapping layer between your application and database. Instead of writing SQL queries, ORM models, and serialization logic, you define a declarative schema and write EdgeQL queries that return exactly the shape of data your application needs. This means fewer lines of code to maintain and fewer opportunities for N+1 query bugs.

The built-in migration system auto-generates migrations from schema changes, so you never write migration files by hand. Run edgedb migration create and the CLI diffs your schema and produces the migration.

How to use

1. Install the CLI: curl --proto '=https' --tlsv1.2 -sSf https://sh.edgedb.com | sh.
2. Initialize a project with edgedb project init, which creates a local instance and schema file.
3. Define your schema in dbschema/default.esdl, create migrations with edgedb migration create, and apply them with edgedb migrate.

Example

# dbschema/default.esdl
module default {
  type User {
    required name: str;
    required email: str {
      constraint exclusive;
    };
    multi posts := .<author[is Post];
  }
  type Post {
    required title: str;
    required body: str;
    required author: User;
    created_at: datetime {
      default := datetime_current();
    };
  }
}

# Fetch users with their posts in one query - no N+1
select User {
  name,
  email,
  posts: { title, created_at }
} filter .name = 'Alice';

Related on TokRepo

• Database tools -- database management and query tools
• AI coding tools -- development tools for building applications

Common pitfalls

• EdgeDB runs on PostgreSQL under the hood but does not expose raw SQL access; if your workflow depends on PostgreSQL-specific extensions (PostGIS, pg_trgm), verify EdgeDB compatibility first.
• The EdgeQL learning curve is real for developers coming from SQL; invest time with the interactive tutorial before building production schemas.
• Self-hosted EdgeDB requires managing a PostgreSQL instance underneath; for simpler deployments, consider EdgeDB Cloud.