PDFs
Find the page. Not just the PDF.
You remember the argument. The diagram. The table. The section that mattered.
You don't remember which report it was in. Or page 47 of which file. Shelf turns supported PDFs into searchable memory.
- 01 · Filemarket-report-final-v3.pdf84 pages. Tells you nothing.
- 02 · PagesPages read on deviceText and structure, kept locally.
- 03 · PassagePage 47: the retention sectionThe part you actually meant.
- 04 · ContextThe pages either side of itMethod, caveats, the footnote.
- 05 · NearbyYour note from that weekKept beside the report.
The decision
You wanted one section. Shelf's job is to return that, with the source still attached.
A filename isn't a memory.
report-final-v3.pdf tells you almost nothing about why you kept it.
What matters is what was inside. Shelf helps make that content retrievable.
- A figure you can describe but not cite
- A paragraph that changed your mind
- A table you compared against something else
- A definition you keep wanting to quote correctly
Remember the idea. Find the page.
Search the meaning you remember. Return to the page where the evidence lived.
Keep the source close to the memory, so quoting it doesn't turn into an afternoon of scrolling.
Four honest document searches.
Nobody remembers filenames. People remember what a page argued.
What you remember
A few pages arguing that onboarding failure looks like churn. It was inside a longer report.
Clues Shelf can use
- Body text about activation and first week
- Sits inside a multi-section report
- Kept while writing the growth review
- Meaning
- The passage is about first-run experience.
- Text inside
- Activation and onboarding appear together on the page.
- Context
- Kept during the growth review.
Kept nearby
- Note: fix week one first · March
- Screenshot: their empty state · March
What does the retrieval work.
Nothing here is document management. It's the same memory engine, pointed at long files.
- 01Semantic searchDescribe the argument; land on the passage.
- 02Contextual searchNarrow by the project and week the document belonged to.
- 03Memory graphSee the notes and captures kept around a report.
- 04RediscoveryOld documents resurface when you're back on the topic.
- 05Research memoryHow a long research trail stays usable.
- 06Local-firstDocuments are read on your Mac, not uploaded.
Documents can contain confidential or personal material.
Contracts, salaries, medical letters, unpublished research, other people's data. Shelf processes supported PDFs on your Mac, which is the only honest way to make a document library searchable.
Nothing is sent to a server to be indexed, and there's no account holding your files.
Documents are useful when the useful part can come back.
Find the part that mattered.
Keep the report. Search the idea. Arrive on the page where the evidence actually lives.
Next thought
Research doesn't get lost in one moment. It gets lost across fifty files, one week at a time.
How to organize researchKeep exploring
Shelf V3 is out now
Shelf V3 is the current build. Shelf does not use automatic updates, so download the latest build here and install it over your current version.
This is intentional. Shelf is built around a local-first, privacy-focused philosophy. An automatic updater requires additional background processes, network connections, update infrastructure, and another system with permission to modify the app. We prefer to keep Shelf’s network footprint and background activity as small as possible.
Manual updates keep the process explicit: Shelf only checks the internet when a feature actually needs it, and you decide when a new version is downloaded and installed.