the one thesis says "Apart from purely visual content, PDF files can also encode a document’s structure. At a language level, Typst has this rich structural information. However, the current implementation does not yet make the best use of it.
Since the frames currently only contain locators and not a full reverse
mapping to the document’s structure, the exporter cannot write this structure
information. As a result, PDFs produced by the Typst compiler are not yet
accessible for visually impaired people." Which bluntly said means that they don't really have a plan here.