Environment Variables

LakeQL does not impose a fixed set of environment variable names. Your lakeql.config.ts decides which variables it reads, typically through a typed env object. Those values are read directly when the config module is evaluated at load time.

1
2
3
4
5
6
7
8
9
10
11
12
// lakeql.config.ts — reads from a typed env module (e.g. @t3-oss/env-core + zod)
import { env } from "./src/env"

export default defineConfig({
  dataSources: {
    warehouse: {
      adapter: "trino",
      connection: { host: env.TRINO_HOST, port: env.TRINO_PORT /* ... */ },
    },
  },
})

A scaffolded project ships an env module and a .env.example file with the variables its chosen adapters need. Copy it to .env and fill it in. Validating the env with a schema makes a missing or malformed value fail fast at load time, with a clear message, instead of surfacing as a connection error later.

Loading the Env File #

• Server — the generated dev/start scripts pass the env file to Node, so it is loaded before the config.
• CLI — pass --env to load an env file before the config is read.

Server Runtime Variables #

Beyond your connection variables, the API server reads its own environment variables. All are optional with defaults:

The CLI reads LOG_LEVEL (default warn) for its own verbosity.