Edit, commit and run

Files

The left side of Build shows your repository’s files. You can:

• Open a file to edit it in the code editor (with syntax highlighting; the editor follows your light/dark theme).
• Create files and folders, rename and delete them (hover a row to see the buttons).
• Upload files or whole folders, either with the upload button or by dragging them in. Drop onto a folder to put them there.

Press Save (or ⌘S / Ctrl+S) to save a file to your working copy.

Commit and pull

Your edits live in a working copy until you commit them. Write a short message and press Commit: the changes go to your GitHub repository. Pull brings in changes someone else made. If the same lines changed in both places, Build tells you which files conflict instead of overwriting anything.

Run dbt

In the Runs panel choose a command and press Run:

Command	What it does
build	Runs models, tests, seeds and snapshots in order. A good default.
run	Builds the models.
test	Runs the data tests.
seed	Loads CSV files from your repo as tables.
compile	Compiles the SQL without running it.
snapshot	Runs snapshots.

• --select (optional) limits what runs, for example my_model+ or tag:daily.
• Source: Working copy runs your files as they are now, including unsaved-to-GitHub changes; Committed code runs a fresh copy of what’s on your branch.

The log streams live. Cancel run stops it. Past runs are listed with their status, so you can open the log of any of them.

Only one run per project is active at a time.

Models in Explore

When a run of committed code succeeds, Explore refreshes its models, so new or changed models, descriptions, dimensions and metrics appear without anyone doing anything. Runs of the working copy never publish, so unfinished work stays private.