Inside HuginnDB
Four screenshots straight out of a running build on Windows, against the Chinook sample database on SQLite. Nothing here is a rendering of what the app might look like.
One workspace per connection. The schema tree on the left — databases → tables → columns with type badges and primary-key markers — and the connection picker in the middle until you open something. The environment rail on the far left keeps separate sets of connections apart.
Double-click a cell — or pick “Open in side editor” — and it opens in a Monaco pane beside the grid, with the value's format detected for you. Format, validate, F11 to go fullscreen, then Save or Discard. The row menu carries the rest: copy with column name, filter by this value, duplicate row.
The same editor, full width, with the schema loaded into autocomplete. Ctrl+Enter runs the statement under the cursor, a Run CodeLens sits above each one, and the results land in a grid underneath with the row count and the timing.
The app resolves the connector binary, then generates the config for whichever AI client you use. Pick the connections to expose, set each one's write policy, copy the snippet. Nothing is exposed until you tick it.
Windows and Linux x86_64 builds ship with every release.
