Marginalias

Marginalias is for reading your own documents more closely — a research PDF, a stack of notes, a long report you need to get through. It brings summarizing and question-answering to that material, but it does the work on your device instead of uploading your files to a server.

You import text-based PDFs, plain text, or Markdown into a local library, and from there you can ask the app to summarize a document, answer questions about it, or pull out key points, action items, outlines, and glossary terms. What sets it apart is that it keeps the output honest: every generated claim is tied to the exact evidence behind it, so you can inspect the source excerpts, character ranges, and PDF pages, and see the retrieved context that produced each answer. When you are studying or triaging, that traceability is the difference between a summary you trust and one you have to re-check by hand.

It is built for focused review, note cleanup, study, and document triage. Output can be copied or shared, the library is searchable, and there is a bundled sample document to try before you import anything of your own. Light, dark, and system appearances are all supported.

Nothing about it requires an account, and there are no analytics or advertising SDKs; your documents and the generated output stay with you. One note on requirements: the AI features run on Apple's on-device intelligence, so they need Apple Intelligence enabled. If it is not available, importing and reading still work, and the app tells you what it needs.