Gel

Gel is a scalable, integrated data platform on top of Postgres. It is built for database development and is described as solving n+1 queries, simplifying migrations, and streamlining database workflow. The service is fully open source.

Its data model uses types rather than tables and includes object types, abstract types, multiple inheritance and mixins, triggers, annotations, introspection, computed fields, user-defined functions, and access policies. Gel supports hierarchical selects, inserts, and updates, link traversal without JOINs, and composable queries. It also provides a strictly typed schema, typed code generation for queries, a TypeScript query builder, EdgeQL and SQL compilers, and SQL or ORM support. The page also describes built-in tools for schema migrations, branching, rapid prototyping, and a graphical UI.

Gel is delivered for both local and cloud use. It includes a complete local development environment with zero configuration, and its CLI can be run with uvx or npx or installed in seconds. Cloud deployments include encryption at rest, backups, monitoring, and export to vanilla Postgres if needed. Other built-in infrastructure includes a connection pooler and cache, built-in server and client connection pooling, and auto-recovery from network and transaction errors in client APIs.

The platform also includes GenAI-related features: a vector store, configurable automatic embeddings, and an automatic RAG endpoint. It lists integrations and SDKs for OpenAI, Anthropic, Mistral, Python, and TypeScript, as well as multiple OAuth integrations, email/password, magic links, passkeys, and a built-in login UI for integrated auth. Additional integrations mentioned include Vercel marketplace billing, GitHub and Vercel previews, and compatibility with most BI solutions.