The hub — every DVT app on this machine, one door.
Cross-engine documentation and lineage: every model, source and column across all your connections, one graph.
Switch models on for REST consumption, control column access, and hand out endpoints instead of database credentials.
Visual development: connections and default target set from a GUI, sources and models dragged into a query editor as source()/ref(), git sidebar, runner, mini-lineage, data preview — and a Build-with-AI pane.
Bring your own LLM key: prompts orchestrate DVT's capabilities on the fly and come back as results and charts you can save for later.
One click adds the profiler package via dvt deps, profiles any supported adapter, and charts the results.
Scheduled dvt jobs — your commands on a cron, with history.
SQL and Python scratchpads over your connections — try things fast, save what's worth keeping.
BI dashboards from basic charts and everything saved out of Chat with Data — multiple reports, one place.