About
We believe researchers should spend time on research,
not file management.
FileMind started as a personal frustration and turned into a product. We built the tool we wished existed — and we're sharing it with researchers who have the same problem.
The problem
Every researcher knows the file. You know the one. It's called
paper_final_FINAL_v3_reviewed.pdf.
It lives in a folder called Downloads, next to 600 other papers with equally
unhelpful names. You know it exists. You cannot find it.
The standard advice is to use a reference manager. So you try Zotero. You import your PDFs. Three hours later, half have wrong metadata, a dozen are duplicates, and you've learned that Zotero works best when papers are added at the time of download — not after years of accumulated chaos.
FileMind is the tool that bridges the gap. Point it at your existing PDF folder. It reads each file, extracts identifiers, looks up real metadata, and proposes a clean, consistent filename. You review and approve. Your library is renamed. Then it builds a semantic search index so you can query your entire library with plain English.
The whole process — for hundreds of papers — takes about 10 minutes. Then you can search, ask questions, and export. No cloud required. No subscription. Just a clean, searchable library on your own machine.
Technical philosophy
Four principles guide every design decision in FileMind.
Local-first
Your PDFs stay on your machine. FileMind runs entirely offline — no cloud sync, no telemetry, no vendor lock-in. Your library is yours.
Evidence-based
Every rename proposal is grounded in actual metadata — DOI lookups, arXiv IDs, extracted title text. We show the source so you can verify.
Non-destructive
Every rename is reversible. FileMind journals each action and can undo batch operations in a single click. Experiment without fear.
Portable
Export to BibTeX, RIS, CSV, or Zotero XML. Your metadata isn't trapped in a proprietary format — take it wherever your workflow leads.
Built on open-source shoulders
FileMind wouldn't exist without the open-source ecosystem. We're grateful to the projects that power the features researchers rely on.
- Ollama — Local LLM inference for semantic search and Ask My Library
- Tesseract OCR — Text extraction from scanned PDFs
- Crossref & Semantic Scholar APIs — DOI resolution and paper metadata
- SQLite — Embedded local database for the search index
- pdf.js — PDF rendering in the preview panel
If FileMind is useful to your research and you work on any of these projects, reach out — we'd love to find ways to give back.
The team
FileMind is built by a small team of researchers and software engineers who got tired of bad filenames. We're currently heads-down on the product. More here soon.
Questions, feedback, or just want to talk about PDF organization? Get in touch.
Ready to reclaim your PDF library?
Free trial — up to 50 PDFs, full feature set. No account required.