Quickstart

Get IronLayer running locally and generate your first execution plan.

Prerequisites

  • Python 3.11+
  • Git
  • A Databricks workspace (for execution — optional for local planning)

1. Install the CLI

install
$ uv pip install ironlayer

Recommended — requires uv and Python 3.11+

2. Initialize your project

Run init inside your SQL project directory. The wizard prompts for your Databricks connection details and creates .ironlayer/config.yaml.

terminal
$ ironlayer init
Databricks host: myworkspace.cloud.databricks.com Warehouse ID: abc123def456 Token: ******** ✓ Config written to .ironlayer/config.yaml

3. Run quality checks (optional)

Validate your SQL models, headers, and project structure with the Rust-powered check engine (90 rules across 12 categories). Runs in under 500ms for 500+ models.

terminal
$ ironlayer check .
Checked 12 files, 4 models 0 errors, 0 warnings ✓ All checks passed

Use --fix to auto-fix fixable issues, or --format json for CI. See CLI Reference for all options.

4. Generate your first plan

Compare the last two commits in the current repo. IronLayer discovers changed SQL models, builds a dependency DAG, and generates a deterministic execution plan.

terminal
$ ironlayer plan . HEAD~1 HEAD
Plan generated plan_id: a7f2c8e1 steps: 4 models: stg_orders, int_revenue, fct_daily_revenue, rpt_revenue strategy: incremental (watermark: 2026-02-20)

5. View the plan

Inspect the plan before executing. See every step, model, and dependency in the execution graph.

terminal
$ ironlayer show plan.json
Step 1 stg_orders CREATE_OR_REPLACE Step 2 int_revenue INCREMENTAL (watermark) Step 3 fct_daily_revenue INCREMENTAL (watermark) Step 4 rpt_revenue CREATE_OR_REPLACE

6. Apply the plan

Execute the plan against your Databricks workspace. Pass --auto-approve to skip the interactive confirmation prompt.

terminal
$ ironlayer apply plan.json --repo . --auto-approve
4/4 models executed successfully total_time: 12.3s cost_est: $0.04 DBU

7. Import from dbt (optional)

If you have an existing dbt project, import it with one command. IronLayer converts dbt models, sources, and materializations into native SQL files.

terminal
$ ironlayer migrate from-dbt ./my-dbt-project
Imported 47 models from dbt project sources: 12 staging: 15 marts: 20

Also supports SQLMesh projects:

terminal
$ ironlayer migrate from-sqlmesh ./my-sqlmesh-project

8. Connect to IronLayer Cloud (optional)

Everything above runs locally with zero cloud dependencies. When you're ready for team features and AI advisory, connect your CLI to IronLayer Cloud.

terminal
$ ironlayer login --api-url https://api.ironlayer.app
Email: aron@example.com Password: ******** Authenticating with https://api.ironlayer.app … ✓ Logged in as Aron Riley Tenant: acme-data Role: admin

Cloud unlocks:

  • AI Advisory — cost predictions, risk scoring, SQL optimization
  • Web Dashboard — plan history, lineage visualization, execution monitoring
  • Scheduled Runs — cron-based execution with automatic retry (coming soon)
  • Team Management — RBAC, shared workspaces

Pricing tiers

Community Team Enterprise
Price Free forever $29/user/month Custom
Seats 1 Up to 10 Unlimited
Plan runs 100/month 1,000/month Unlimited
AI advisory calls 500/month 5,000/month Unlimited
Models 5 Unlimited Unlimited
API requests 10K/month 100K/month Unlimited

See Pricing for the full comparison.

Next steps