LakeQL is a monorepo that provides a type-safe GraphQL access layer for Trino-powered data platforms. It combines a backend runtime with a CLI that generates schema and query files from your existing data model.
The goal is to reduce manual boilerplate while keeping your API predictable, secure, and easy to extend.
- Type-safe GraphQL APIs for Trino-powered data platforms
- CLI-based schema generation from existing Trino metadata and table structures
- Built-in filtering, sorting, and pagination for generated queries
- Extensible response transformation with reusable helper packages
- Trino authentication with Basic Auth and OAuth flows
- Monorepo setup with reusable packages and a ready-to-use app template
- Node.js 24
- pnpm 11+
Run the following command in your terminal:
pnpm dlx @lakeql/create-appThen follow the interactive setup.
The command bootstraps a project from the preconfigured template in templates/app.
Run the interactive CLI to generate your schemas.
pnpm cli pull --source-path ./srcTo start local development, run:
pnpm devThe template ships with a production-ready tsdown setup.
Use pnpm build to generate the production dist output, which you can then use in your deployment setup (for example, in a Dockerfile).
Full docs and examples are available at https://lakeql.dev
- Local Development — set up minitrino, seed test data, run queries
- Contribution Guide — code style, changesets, project structure
This project has its roots in the original datalake-graphql-wrapper.
LakeQL was later rebuilt from the ground up as a fully type-safe monorepo, with a complete refactoring of the codebase and architecture.
- Igal Klebanov and the Kysely Team - Kysely is used to generate valid SQL statements from GraphQL queries, and Igal helped make the query builder more generic.
- Michael Hayes - He created Pothos GraphQL, which is the foundation of our GraphQL server and its code-first schema generation approach.