generated: '2026-09-05'
method: searched
source: >-
  https://docs.databricks.com/aws/en/getting-started/free-edition ;
  https://www.databricks.com/product/pricing
summary: >-
  There is no bundle-specific sandbox or mock deployment target. What a
  developer can get for free is a real Databricks workspace — Free Edition —
  and deploy real bundles into it. The rehearsal mechanism for bundles is
  `databricks bundle plan` / `validate`, not a test-mode credential.
environments:
  - name: Databricks Free Edition
    type: free-tier-workspace
    url: https://docs.databricks.com/aws/en/getting-started/free-edition
    trial: false
    permanent: true
    description: >-
      "a no-cost version of Databricks designed for students, educators,
      hobbyists, and anyone interested in learning or experimenting." Serverless
      only and quota limited.
    includes:
      - Build AI applications and agents
      - Collaborate on data science and ML projects with shared notebooks
      - Explore and analyze real datasets via SQL editor or notebooks
      - Create interactive dashboards using Genie
      - Design and test data pipelines through Lakeflow
      - Genie Code in-product coding assistance
      - Lakebase Postgres
    limits: >-
      Described as "serverless-only, quota-limited"; the specific quota numbers
      live on a separate Free Edition limitations page and are not restated
      here.
  - name: Databricks free trial
    type: trial
    url: https://www.databricks.com/try-databricks
    description: >-
      Free access to the Databricks Data + AI Platform. The pricing page does
      not state the trial length, so none is recorded.
    caveat: >-
      "If you configure Databricks to work with your own cloud account, you will
      still be charged by your cloud provider for resources."
test_credentials:
  published: false
  note: >-
    Databricks publishes no test-mode API keys, no test key prefix, and no
    fixture identifiers. Bundles authenticate with the same OAuth credentials
    against a real workspace regardless of target; separation is by workspace
    and by deployment mode, not by a test key.
rehearsal:
  - command: databricks bundle validate
    description: Validate configuration syntax without deploying.
  - command: databricks bundle plan
    description: Show the deployment plan without making changes.
  - command: databricks bundle sync --dry-run
    description: Report which files would sync without writing them.
  - mechanism: development deployment mode
    description: >-
      Prefixes every resource with "[dev <user>]", pauses all schedules and
      triggers, and tags resources dev — the closest thing to a sandbox the
      product ships, inside a real workspace.