CLI Reference

The IronLayer CLI (ironlayer) provides commands for planning, executing, and managing your transformation workflow.

Project Setup

`ironlayer init`

Initialize a new IronLayer project. Creates a config file, example models, and .env.

ironlayer init [--non-interactive] [--name NAME] [--state-store local|postgres] [--no-ai] [--no-git]

`ironlayer dev`

Start a local development server with SQLite state store and DuckDB execution.

ironlayer dev [--port 8000] [--ai-port 8001] [--no-ai] [--no-ui] [--reload]

Authentication

`ironlayer login`

Authenticate with an IronLayer API instance. Stores credentials in ~/.ironlayer/credentials.json.

ironlayer login --api-url https://api.ironlayer.app --email you@company.com

Flag	Description
`--api-url`	API base URL (prompted if not provided)
`--email`	Login email (prompted if not provided)
`--password`	Account password (prompted securely if not provided)

`ironlayer logout`

Remove stored credentials.

`ironlayer whoami`

Show the current authenticated user, tenant, and role.

Planning

`ironlayer plan`

Generate a deterministic execution plan from a git diff.

ironlayer plan <repo_path> <base_ref> <target_ref>

Argument / Flag	Description
`repo_path`	Path to the git repository containing SQL models
`base_ref`	Base git ref (e.g. `HEAD~1`)
`target_ref`	Target git ref (e.g. `HEAD`)
`--out, -o`	Output path for plan JSON (default: `plan.json`)
`--as-of-date`	Reference date for date arithmetic (YYYY-MM-DD)

`ironlayer show`

Display a human-readable summary of a plan file.

ironlayer show <plan_path>

Argument	Description
`plan_path`	Path to the plan JSON file (e.g. `plan.json`)

Execution

`ironlayer apply`

Execute a previously generated plan against your Databricks workspace (or DuckDB locally).

ironlayer apply <plan_path> --repo <repo_path> [--auto-approve] [--override-cluster <name>]

Argument / Flag	Description
`plan_path`	Path to the plan JSON file
`--repo`	Path to the git repository (required)
`--auto-approve`	Skip manual approval prompt
`--approve-by`	Name of the person approving this execution
`--override-cluster`	Override the cluster/warehouse used for execution

Backfills

`ironlayer backfill`

Run a targeted backfill for a single model over a date range.

ironlayer backfill --model <name> --start 2026-01-01 --end 2026-02-01 --repo <path>

Flag	Description
`--model, -m`	Canonical model name to backfill
`--start`	Start date (YYYY-MM-DD, inclusive)
`--end`	End date (YYYY-MM-DD, inclusive)
`--repo`	Path to the git repository (required)
`--cluster`	Override cluster/warehouse

Discovery

`ironlayer models`

List all SQL models discovered in a repository.

ironlayer models <repo_path>

`ironlayer lineage`

Display upstream and downstream lineage for a model. Use --column for column-level tracing.

ironlayer lineage <repo_path> --model <name> [--column <col>] [--depth N]

Argument / Flag	Description
`repo_path`	Path to the repository containing SQL models
`--model, -m`	Canonical model name to trace lineage for
`--column, -c`	Column name for column-level lineage (optional)
`--depth`	Maximum traversal depth (default: 50)

Migration

`ironlayer migrate from-dbt`

Import models from an existing dbt project into IronLayer format.

ironlayer migrate from-dbt <project_path>

`ironlayer migrate from-sqlmesh`

Import models from a SQLMesh project.

ironlayer migrate from-sqlmesh <project_path>

`ironlayer migrate from-sql`

Import plain SQL files.

ironlayer migrate from-sql <sql_directory>

MCP Server

`ironlayer mcp serve`

Start an MCP (Model Context Protocol) server for AI assistant integration.

ironlayer mcp serve [--transport stdio|sse] [--host 127.0.0.1] [--port 3001]

Global Flags

Flag	Description
`--json`	Output machine-readable JSON to stdout
`--metrics-file`	Write execution metrics to a JSON file
`--env`	Target environment (default: `dev`)

Environment Variables

Variable	Description	Default
`IRONLAYER_API_URL`	API base URL	Stored from `login`
`IRONLAYER_API_TOKEN`	Auth token (overrides stored credentials)	—