Skip to the content.

Home · Quick start · Editor · charter.yaml · Sources · Agent · Guides · Evals · Audit · Policies · CLI · MCP · Workspace · Desktop · About · FAQ

The workspace

A DataCharter workspace is just a directory. Everything that describes your exploration environment lives in files you can read, diff, commit, and share — and nothing secret or machine-specific travels with them.

What’s in it

my-workspace/
  charter.yaml        # sources, tables, PII fields — your catalog as a contract
  queries/*.sql       # saved queries (the Query Files panel reads/writes these)
  .env.example        # placeholder secret names, committed
  .env                # real secrets — git-ignored, never committed
  .datacharter/       # local state (cache, snapshots, temp) — git-ignored

datacharter init scaffolds the committable parts; datacharter init --demo adds a small generated dataset so you can try everything immediately.

Snapshots — save a result as a local table

Run a query, click Snapshot, and the result is saved as local.<name> — a reusable table kept in .datacharter/ on your machine. Query it like any other relation; datacharter recheck <name> re-runs its SQL and diffs against the saved copy. To remove a snapshot (or a table you dragged in) later, click the next to it in the source tree.

Snapshot a query result as local.top_spend, then query it

Portable by construction

The result: your team’s whole data-exploration setup is a repo. Clone it, run it, and you’re looking at the same sources — while credentials and local state stay on each person’s own machine.