Current behavior: Wiki imports currently support .zip, .md, and Markdown vaults such as Obsidian or Notion exports. User problem: A lot of important reference material lives outside Markdown. For example, university course syllabi are usually PDFs and contain key information such as exam dates, weekly content, assignments, required readings, and course policies. Right now, these files need to be re-uploaded every time an AI conversation needs to reference them, creating enough friction that users may avoid using the feature altogether. Desired behavior: Users should be able to import PDFs and images into a wiki once, then reference them persistently across future conversations. Example workflow: "Check my Calculo I syllabus on Membase." The AI retrieves the syllabus from the user's persistent wiki without requiring a new upload. Why it matters: PDF support is especially important because so much useful reference material exists in that format - academic papers, course syllabi, technical specs, contracts, school documents, and product documentation. This would make Membase feel more like a true persistent knowledge base.
