Private research stack on Mac, from figure to finding in one shortcut

Research time disappears into scrolling. You remember a figure caption, a method name, or a result threshold, not the filename. The goal is simple, type what you remember and land on the exact page in the right paper so you can cite with confidence.

Fenn gives you a private, on device research stack for macOS. It searches inside PDFs, slides, images, and even recordings you keep with your notes, then opens the exact page, slide, frame, or timestamp. Add your libraries once, use one shortcut, and move from memory to evidence in seconds.

Why papers slow you down

• Filenames and folders do not reflect figures, tables, or methods.

• Exported PDFs hide the one panel or equation you need.

• Finder and Spotlight return files, not the right page.

• Image heavy figures and scanned papers make text search unreliable.

What Fenn does for researchers

• Search by content, not name. Papers, supplementary PDFs, slide decks, screenshots, and scanned PDFs.

• Open the exact moment. Page numbers for PDFs, slide numbers for decks, text within images, timestamps in recordings.

• On device by default. Indexing and search run locally on your Mac for privacy.

• Four modes to match how you think. Semantic, Keyword, Hybrid when you want both, and Exact for precise matches.

• Agent mode for complex questions. Ask criteria based queries that combine entities, dates, thresholds, or proximity rules. Fenn analyzes your local files and returns context rich results you can open immediately.

From figure to finding in one shortcut

1. Choose sources once
   Add folders where you keep papers and notes, downloads from publishers, supplementary materials, slide exports, and screenshots.

2. Index on your Mac
   Fenn makes the text inside PDFs and images searchable. Your files stay local.

3. Use the keyboard shortcut
   Open Fenn anywhere on macOS. Pick Semantic, Keyword, Hybrid, or Exact.

4. Type what you remember
   Try these researcher friendly prompts:

   • figure caption “Kaplan–Meier”

   • table “ablation study” accuracy

   • method “LoRA fine tuning”

   • “p < 0.05” near “AUC”

   • “Figure 3” latency distribution

5. Open at the exact page
   Press Return to land on the precise page in your PDF viewer. Copy the quote with page context and continue.

6. Use Agent mode for criteria based recall
   Examples you can run:

   • papers with “ROC AUC” above 0.90 in 2024 mentioning “breast MRI”

   • methods sections that cite “AdamW” and “weight decay 0.01”

   • figures where “latency P95” exceeds 200 ms in mobile benchmarks

   • tables including “parameter count” within 10 words of “billion”

Real world research use cases

Literature review
Search “Figure 2” ablation or a method keyword and jump straight to the right panel in multiple papers. Collect images and citations with page references as you go.

Reproduction and benchmarking
Query “batch size 32” near “learning rate” or optimizer AdamW and open methods pages to align settings quickly.

Clinical or scientific evidence packs
Ask trials with N > 500 and p < 0.01 or a disease plus modality. Open the exact table that confirms inclusion criteria.

Talk prep
Search “failure cases” or “limitations” and jump to slides or pages that inform your discussion without scanning whole PDFs.

Mini case, before and after

Before
You spend 25 minutes skimming four PDFs to find one figure, one table, and one quote for a slide. Window flipping and Ctrl-F guesses waste the block.

After with Fenn
You run three queries, open three pages exactly where needed, and paste with page references. Total time about 5 minutes. The rest of the block goes to analysis, not hunting.

Privacy matters for research

Grant submissions, preprints under review, clinical data, and lab notes cannot leave your machine. Fenn is on device by default. Your files remain on your Mac during indexing and search. You get fast recall and precise opens without sending sensitive material to external services.

Tips for cleaner, faster recall

• Use Hybrid mode when you know a keyword and the idea behind it, for example “contrastive loss” plus “augmentation”.

• Use Exact mode for citations or numeric thresholds you want to match precisely.

• Keep supplementary PDFs and exported figures in the same project folders you index.

• For scanned papers or figure images, include a nearby labeled term, for example “Kaplan–Meier”, “95% CI”, “confusion matrix”.

Setup

• Add ~/Downloads, your literature folder, and your notes or slides folder as sources.

• Let Fenn index locally.

• Map the keyboard shortcut.

• Run two or three queries from your current project and open the exact pages to prove the loop.

Keep exploring

• Find any line in a PDF on Mac, instant page jumps for long papers

• Search text inside images and screenshots on Mac, useful for figure panels

• Best private AI tools for macOS in 2025, privacy first picks

Search by caption, metric, or method. Open the exact page. Keep your work private. With Fenn on macOS Apple Silicon, you can find the moment, not the file.