generated: '2026-07-18'
method: searched
source: https://dlthub.com/docs/reference/command-line-interface
name: dlt
summary: >-
  First-party command line interface shipped with the dlt Python library for
  scaffolding, inspecting, deploying, and operating data pipelines.
install:
  - method: pip
    command: pip install dlt
  - method: pip-extras
    command: pip install "dlt[cli]"
    note: deployment features (dlt deploy)
commands:
  - name: dlt init
    purpose: Create a pipeline in the current folder from an existing verified source or a new template.
  - name: dlt pipeline
    purpose: Inspect pipeline state, trace, and load packages; provides basic maintenance.
    subcommands:
      - { name: info, purpose: Display pipeline state and working directory contents }
      - { name: show, purpose: Launch the workspace dashboard with load status and data explorer }
      - { name: failed-jobs, purpose: Show failed loads and their error messages }
      - { name: drop-pending-packages, purpose: Delete extracted and normalized packages }
      - { name: sync, purpose: Reset local pipeline state while preserving destination data }
      - { name: trace, purpose: Show the last run trace with timing and step details }
      - { name: schema, purpose: Display the default pipeline schema }
      - { name: drop, purpose: Selectively remove tables and reset associated state }
      - { name: load-package, purpose: Display information on a specific load package }
      - { name: mcp, purpose: Launch an MCP server attached to this pipeline }
  - name: dlt schema
    purpose: Show, convert, and upgrade schemas.
  - name: dlt deploy
    purpose: Create a deployment package for a selected pipeline script.
    subcommands:
      - { name: github-action, purpose: Deploy the pipeline to GitHub Actions with cron scheduling }
      - { name: airflow-composer, purpose: Deploy the pipeline to Google Cloud Composer / Airflow }
  - name: dlt telemetry
    purpose: Show telemetry status.
  - name: dlt dashboard
    purpose: Show the dltHub workspace dashboard.
  - name: dlthub ai
    purpose: AI / agent tooling (formerly `dlt ai`); requires `pip install dlt[hub]`.