PaperHoundThe papers you cite, downloaded and filed.
Paste a reference list. PaperHound works out which papers they are, downloads the open-access PDFs, opens the rest through your university library, and files everything into Mendeley or Zotero. Then it turns any PDF into Word, Excel or PowerPoint.
Find a paper: one search, every source at once.
From a reference list to a filed library.
Hours of hunting become a few minutes. Here is the whole of it, area by area.
Drop a file, paste a list, or search.
- Drop a file with links: Word, Excel, PDF, text, CSV or HTML. Every link in it is found and the PDFs downloaded to a folder you choose.
- Paste references: links, DOIs or plain citations in any style, one per line. Citations are identified through Crossref and kept only when the match is certain. Books are recorded as references.
- Find a paper: title words, an author, a year, in any mix. Every source is asked at once; filter by kind, years, citations, region and source; sort by best match, newest, oldest or most cited.
- Tags from spreadsheets: columns named theme, chapter, priority or topic become tags that travel into the PDFs, the RIS and BibTeX files, Mendeley, Zotero and the Word summary.
- Watch a file: links added while the app is open are picked up by themselves.
- Reference Expert: Apple's on-device AI reads references the normal reader can't settle, in any style or language. Its answers are checked against the registries first, and nothing is sent to an AI service.
Quick first, then as deep as it takes.
- Quick Download asks the registries and direct links. Deep Download adds archives and page reading for anything that failed. Stop and resume at any time.
- Deep Retry goes after what's left: Internet Archive copies, open-access search by title, deeper crawling, longer waits.
- Reassess rereads the whole list with the AI to catch references the first pass missed.
- Checked, not guessed: a link has to serve the cited paper before the PDF is kept. Duplicates are spotted; PDFs you downloaded yourself can be added by hand.
- Every PDF stamped with its title, authors, subject and tags, so it's findable later.
- Cited-by count and retraction warning on every paper, from OpenAlex.
Paywalled papers, through your own library.
A minimal, one-time setup connects PaperHound to the access your university already pays for.
Library proxy (EZproxy)
Paste your institution's proxy prefix once. Paywalled papers then get "Open via Library Proxy", you sign in in your browser, and the app never sees your password. Extra publisher domains can be routed through the proxy too.
Library search and full-text links
Add your discovery site (Primo, EDS, Summon and others) and every paper gets "Open in Library Search". With your OpenURL resolver, papers with a DOI open straight on the full-text page.
Library Browser
A built-in browser walks you through the papers that failed, one at a time, and captures each PDF automatically or with one click. It remembers your library sign-in, keeps a count of captured, skipped and remaining, and can search Google Scholar or DuckDuckGo for the paper in front of you.
The ones that are left
Whatever still hasn't downloaded is grouped by reason, with Open All in your browser, through the proxy or in library search, and Copy Links, so nothing is lost.
Filed where you work.
Send to Mendeley
References go in with their PDFs attached and the right details, duplicates removed, through your own Mendeley developer key. A result sheet lists anything that couldn't be attached.
Send to Zotero
Through the Zotero connector on the same computer, by hand or automatically after each run.
RIS and BibTeX
Both files are written after every run, for EndNote and anything else. Copy the whole list, or a single citation, with one click.
Report and Word summary
A download report with citations, cited-by counts and retraction flags, and a Word document summarising the run, both opened in a click.
Any PDF into Word, Excel or PowerPoint.
- PDF to Word: exact layout, so pages look as they do, or flowing text, so paragraphs reflow and edit like a normal document.
- PDF to Excel: every table in the PDF, one sheet per table, with figures kept as numbers.
- PDF to PowerPoint: one editable slide per page, with text, tables and pictures in place.
- Choose the pages from a preview, with ranges like 1-3, 7-. Convert many PDFs at once, including every PDF a run downloaded.
- Scanned pages are read with on-device text recognition, with a second reader for figures and lines the first is unsure of. Multi-column pages, equations and embedded fonts are handled.
- Nothing leaves the machine. Both readers run on the device.
Mac, Windows, iPhone and iPad.
Mac
The full app: downloading, Find a paper, Reference Expert, library access, Mendeley and Zotero, PDF conversion. Settings, keys, library addresses and history sync to your iPhone and iPad through iCloud, with no account.
iPhone and iPad
The same app, with folders through the Files app and converted files shared to Mail, Word or anywhere. The Library Browser works on the phone too.
Windows
Downloading from the registries and archives, the library proxy and Library Browser (which remembers your sign-in), sending to Mendeley and Zotero, the report and Word summary, and optional account sync. Find a paper, Reference Expert and PDF conversion are Mac and iOS only for now.
Privacy. Only lookups (titles, DOIs, links) go to the registries and search engines. References and PDFs go to Mendeley or Zotero only when you send them. The AI and the text recognition run on your device. Your library password is typed in your browser, never in the app. PaperHound uses open-access sources and your own library's access; it never uses pirate sites.
Things researchers ask first.
What can I paste?
Anything: web links, DOIs, or plain citations in any style ("Salomon and Murray (2002). The epidemiologic transition revisited. Population and Development Review."), one per line. You can also drop a Word, Excel, PDF, text, CSV or HTML file and the links inside it are found.
How does it get into my university library?
You paste your institution's proxy prefix (the ezproxy address on your library's website) once in Settings. From then on, paywalled papers open through the library; you sign in in your browser the normal way. Adding your library's search page and OpenURL resolver gives one-click full-text links too.
Does it use Sci-Hub or similar?
No, and it never will. It uses the open-access registries and archives listed above, and your own library's legitimate access.
Do I need any API keys?
No. Everything works without them. Optional keys for Semantic Scholar, CORE and Brave Search raise the limits and improve the web search; Settings has a button to test them.
How does Mendeley sending work?
Through Mendeley's own API with a client ID and secret you create free at dev.mendeley.com (a few clicks, once). References arrive with their PDFs attached and duplicates removed. Zotero works through its connector on the same computer, and RIS and BibTeX files cover everything else.
How good is the PDF to Word conversion?
Exact layout keeps pages looking as they do, including multi-column text, tables, equations and embedded fonts; flowing text gives you a normal, editable document. Scanned PDFs are read with on-device text recognition. Tables go to Excel as real numbers, and pages to PowerPoint as editable slides.
How do I get it?
PaperHound is in testing. Email support@mabeleinnovations.com and say which platform you use; we'll send you the TestFlight link (Mac, iPhone, iPad) or the Windows installer.
Stop hunting for PDFs.
Join the test, paste your next reference list, and watch it fill your Mendeley library.