Skip to content

Transformations as pipelines

Describe a clean-up or reshape in plain language; canTrust previews the generated code on a safe clone and snapshots before applying.

AI-Assisted Data Pipelines

canTrust's Easy ETL feature lets you clean, transform, and load data using natural language instructions. Describe what you want — "remove duplicate rows", "convert dates to ISO format", "merge these two tables on customer ID" — and canTrust generates and executes the pipeline for you.

How It Works

  1. Describe the transformation — tell the AI what you need in plain English.
  2. Review the plan — canTrust generates SQL or Python code and shows a preview of the changes on a safe in-memory clone.
  3. Execute with confidence — approve the transformation. canTrust snapshots your data before applying changes, so you can always roll back.
  4. Validate results — the AI shows before/after comparisons and impact estimates (rows affected, tables modified).

Materialization

Transformed data is materialized into a local DuckDB or SQLite store for fast, repeatable querying. This means your original data source is never modified — canTrust works on a local copy that stays in sync with your connection settings.