Reports
A report is generated from a conversation on demand. You do not assemble it by hand from scratch: the chapters come out of what was calculated, and you edit from there.
Ask for it in the chat when you are happy with the result:
Put this into a report: the load assumptions, the verification, and the deflection check.
Add the cross-section table as an appendix.
You get a document with chapters, and a PDF you can download.
The editor
The report panel shows the chapters as a document you can edit directly. Rewrite a paragraph, reorder chapters, drop the ones you do not want, add a heading. What you see is what the PDF contains.
One thing worth knowing about how it stores your text: the chapter content is Markdown, and that Markdown is the single stored truth. The chat view, the editor, and the PDF all render the same bytes. If a heading or a table looks different in the PDF than in the editor, that is a rendering bug worth reporting, not a conversion you have to work around.
Example: you paste a table into a chapter and it looks right in the editor but breaks in the PDF. Do not reformat it by hand until it survives both. Report it. The three surfaces are supposed to agree, and the pipeline has tests that hold them byte-exact.
The cover page
The cover carries your company logo, the project, the planner, the client, and a date. It is filled from the project data, so setting a project up once means every report from it starts correct.
Enterprise plans can go further and suppress the Dlubal footer badge entirely, with a custom footer text and accent colour, so the document looks like your office produced it.
What it costs
A PDF export costs 2 credits. Editing chapters costs nothing, so iterate in the editor and export once you are done rather than exporting after every change.
PDF reports are not on the Free plan. Everything else on this page (the chapters, the editor) is about producing them, so on Free you can calculate and read results but not generate the document.
Fonts and languages
Reports render in the language of your conversation. You do not switch a report language separately.
Script coverage is not complete yet, and it is better that you know which side of the line your language is on:
| Script | In a PDF today |
|---|---|
| Latin, including diacritics (German, Czech, Polish, Turkish, …) | Renders correctly |
| Chinese | Renders correctly, through an embedded fallback font |
| Japanese | Kana renders today through the same font, but Japanese is not wired up as its own case yet |
| Korean, Russian, Ukrainian, Hindi | Not covered yet, characters come out as empty boxes |
| Arabic | Deferred, the PDF engine cannot shape right-to-left text |
Example: a conversation held in Czech produces a Czech report with the diacritics intact. The same conversation held in Russian produces a PDF where the Cyrillic text is boxes. If your language is in the bottom two rows, read the results in the chat and hold off on the PDF.