Gel (formerly EdgeDB) Graph-Relational Database on PostgreSQL

This is a repackaged open source software product wherein additional charges apply for cloudimg support services.

Backend engineers and data-modelling teams running workloads on AWS EC2 often face a choice: spend 30-60 minutes manually installing a database, configuring TLS, setting up a reverse proxy, and generating credentials - or use a managed service that limits control. This Gel AMI eliminates that trade-off. You get a fully configured graph-relational database that is query-ready on first boot, with production-grade security defaults and full administrative access.

Gel (formerly EdgeDB) is an open source graph-relational database built on top of PostgreSQL. You describe your data as typed object types with links and properties in a declarative schema, evolve it through versioned migrations, and query it with EdgeQL - a composable query language that returns rich, deeply nested results without hand-written joins. Gel ships a built-in web admin interface (the Gel UI) with a schema browser, an interactive REPL, and a data explorer. Official client libraries cover popular languages.

• Boot to first query in minutes: Gel is installed from the official package repository and runs under systemd behind an nginx reverse proxy. No manual package installation, no proxy configuration, no credential setup.
• Demo schema pre-seeded: A small schema and data set are loaded via a recorded migration on first boot, so the schema browser and data explorer are populated the moment you sign in.
• Paired deployment guide included: Walk through reading your credentials, exploring the schema, running EdgeQL, applying a migration, and connecting a client library.

Compare this to a bare EC2 install where you would manually add the Gel repository, configure systemd units, generate TLS certificates, set up nginx, and create authentication credentials - steps that typically take 30-60 minutes and introduce configuration drift across instances.

Unlike images that ship shared credentials, this AMI generates all secrets fresh on every first boot:

• Unique superuser password - randomly generated per instance, never baked into the image
• Unique self-signed TLS certificate - created during data directory initialisation
• Unique nginx reverse-proxy password - no two instances share authentication material

The database process binds only to the loopback interface and is reachable solely through the authenticated nginx proxy on port 80. systemd manages the service for automatic restarts and clean logging. The image is captured with the data directory wiped and the server disabled, ensuring every launch starts clean.

Consider a SaaS product catalog with deeply nested categories, variant options, and cross-linked recommendations. In raw PostgreSQL you would write multi-level JOINs or recursive CTEs. With Gel, you model categories, products, and recommendations as typed object types with links, then retrieve an entire nested product tree in a single EdgeQL query - no ORM layer, no manual join assembly. The Gel UI lets you visually browse that schema and test queries interactively before wiring them into your Next.js or Python backend via the official client library.

• Gel 7 (current release) installed from the official repository
• nginx reverse proxy with HTTP basic authentication on port 80
• Self-signed TLS between proxy and database
• systemd service management with automatic restart
• Pre-seeded demo schema and data set
• Deployment guide covering credentials, schema exploration, EdgeQL, migrations, and client connection
• 24/7 cloudimg support via email and live chat

Launch the AMI on your preferred EC2 instance type, wait for first-boot initialisation to complete, then retrieve your generated credentials from the instance. Sign into the Gel UI through your browser, explore the demo schema, run EdgeQL queries in the REPL, and connect your application using an official client library. The deployment guide walks you through each step.

Gel and EdgeDB are trademarks of their respective owner. All product and company names are trademarks or registered trademarks of their respective holders. Use of them does not imply any affiliation with or endorsement by them.