Paywalled papers
Recipes for papers you can read through a subscription or an account but that have no open copy.
An Elsevier paper
mosaic get 10.1016/j.paerosci.2006.01.001 --auto--auto drives your own Chrome (or Edge; on WSL the Windows one) through the paper's PDF chain and files each download it saves; ScienceDirect's fingerprint wall passes because the driver is a genuine browser. Where the mosaic profile has no login, the paper is skipped as "login needed"; sign in there once in the mosaic window and re-run. Guide: auto pass.
Through your institution's proxy
mosaic config --proxy-url 'https://biblioproxy.cnr.it:2481/{doi}'
mosaic get --from refs.bib --autoCopy your proxy's address from a paper you opened through it, with {doi} in place of the DOI, and sign in to the proxy once in the mosaic Chrome window. Publishers the proxy covers then download with your institution's subscription; when the proxy session ends mid-run, mosaic opens its sign-in page and goes on once you are in. Guide: institutional proxy.
The same paper, you click
mosaic get 10.1016/j.paerosci.2006.01.001 --handoffWithout Chrome, or when a site shows a check a browser must pass by hand: mosaic opens the article page in your default browser and watches your Downloads folder. Click "Download PDF" as usual; mosaic checks the new file is a PDF, moves it into the library under the paper's name and stops. Tutorial: your browser does the download. Guide: hand-off.
Every failed paper of a reference list
mosaic get --from refs.bib --auto
mosaic get --from refs.bib --auto --interactive # the auto pass, then a window for the rest
mosaic get --from refs.bib --handoff # you click, mosaic filesThe open papers download as usual; the subscribed ones run through the auto pass, one publisher after the other. A paper behind a login you do not have is skipped, not waited on; a Cloudflare check waits up to 3 minutes for your click. --max-auto N raises the 50-paper per-run cap for the auto pass. Each paper's ledger is in the cache, so re-running the same command resumes where it stopped.
mosaic cannot find your Downloads folder
mosaic get 10.1016/j.paerosci.2006.01.001 --handoff --watch-dir /mnt/c/Users/me/DownloadsOn WSL mosaic asks Windows where the Downloads folder is, also when it was moved into OneDrive; on Linux it asks xdg-user-dir DOWNLOAD, then tries ~/Downloads. --watch-dir names the folder when neither works, or when your browser saves elsewhere.
PDFs you already downloaded
mosaic import ~/Downloads/1-s2.0-S0094576517312663-main.pdf
mosaic import ~/Downloads/*.pdf --download-dir ~/papers/hall-thrustersimport finds each file's DOI (the PDF's metadata, then the article ID in a ScienceDirect file name, then a single DOI in the text), takes title, authors and year from the cache or Crossref, and moves the file into the library under the usual name. Files that are not PDFs, or whose DOI it cannot find, are reported and left in place; the command exits with status 1 if any file failed. Tutorial: PDFs you already have. Guide: PDFs you downloaded yourself.
A PDF without DOI metadata
mosaic import scan.pdf --doi 10.1016/j.paerosci.2006.01.001--doi names the paper when detection fails, and works with one file at a time.
Replace a file already in the library
mosaic import ~/Downloads/better-copy.pdf --doi 10.1016/j.paerosci.2006.01.001 --forceWithout --force, import refuses to overwrite a PDF already filed under the same name.
Log in to a publisher once
mosaic auth login springer --url https://link.springer.comA browser window opens at the URL. Log in as the site asks (institution, SSO, two-factor), wait until the publisher's page shows you logged in, then press Enter in the terminal: mosaic saves the cookies to ~/.config/mosaic/sessions/springer.json. Later downloads use the session without any flag. Pressing Enter in the middle of SSO redirects saves a session that does not work. Needs the browser extra and a Playwright browser (Installation). Guide: saved sessions.
Check your saved sessions
mosaic auth statusThe Login column tells how long ago a session was saved: publisher logins live in cookies that end within hours or days, so a session saved more than a day ago is probably logged out. Log in again, or let a successful get --interactive refresh it.
A CAPTCHA or an SSO prompt
mosaic get 10.1088/1361-6595/ab46c5 --interactive--interactive (-i) retries the failed papers in a visible browser window loaded with your sessions. When the PDF does not arrive by itself, the terminal says it is your turn: log in, solve the check or open the PDF in the window. It needs a display; on WSL2, WSLg provides one. Guide: interactive download.
Forget a session
mosaic auth logout springerDeletes the saved session files for that name.