Guide

Everything Link Meteor does, where to find it, and what to expect. New here? Start with installation, then try the practice page.

Where things are

On the page: While you select a region, a small bar at the top of the page explains the gesture, matching links light up, and a count follows your pointer. When you let go, a card lists what you selected and where it will be saved.
Side panel: Click the Link Meteor icon in Chrome's toolbar to open the side panel beside the page you're reading. Click the icon again to close it. It shows your active collection, capture options and your links. Use the folder button for collections and site settings, and Export at the bottom for copying and downloading.
Full view: The Full view button at the top of the side panel opens the same workbench in its own tab, with your collections, links and export options side by side, and links shown as anchor text, URL and source columns. The side panel closes when it opens. Open the full view is also in the toolbar icon's right-click menu.
About and help: The ? button shows the version and links to this guide and to bug reports. It also credits PDF.js by Mozilla, which is part of Link Meteor and reads PDFs in your browser. Copy diagnostics there copies versions, settings, permission choices and counts to paste into a bug report. It never includes addresses, site names, page titles, notes, tags or collection names.

Select a region

Press ⌥ ⇧ L on a Mac or Alt Shift L on Windows and Linux. You can also choose Select a region in the side panel, or right-click the page and choose Link Meteor, then Select a region. Then drag a rectangle across the links you want.

A link is included when any visible part of it overlaps your rectangle, even by a pixel. Links that wrap across several lines count if any line overlaps.
Drag toward the edge of the window, or use your scroll wheel while dragging, to scroll the page or a scrolling box. Links you've already passed stay selected, and their highlights stay on them as the page scrolls.
Press Esc at any time to cancel. Nothing is saved until you choose an action.

When you let go, choose one of:

Copy text + URL: Copies two tab-separated columns, anchor text and URL, with a header row, ready to paste into a spreadsheet. Nothing is saved.
Add to collection: Saves the links to the collection named on the card.
Review: Saves them and opens the full workbench.
Add another region: Saves these links, then starts a new selection so you can keep going.
Open in tabs: Opens the web links as background tabs. See opening links.
More: Open in a new window or as a tab group, copy the URLs, Markdown links or rich links, download the selection as an Excel workbook, download the files behind the links, or save it as a bookmark folder.

The card helps you choose:

Filters: All, Other sites, PDFs and Same site tick only the links that match.
Already saved: links the collection already has are marked Saved. Skip saved leaves them out when adding.
Navigation links: with Leave out navigation links on, links in a page's menus, header, footer and sidebars start unticked, and the card says how many. Include them ticks them again.

The card lists the links it found. Untick any you don't want, and every action leaves them out. The Adds to line names the collection the links will go to; choose it to pick another. While the card has focus, each action has a one-letter shortcut, shown in its tooltip: C copies, A adds, O opens, U copies URLs, K copies Markdown, L copies rich links and F downloads the files, for example. Unticked links keep a dashed outline on the page.

Hold-key drag

Skip the shortcut: hold a key and drag across links on the page. The settings are under Site access, in the side panel's collections and site settings.

On every site: Choose Allow on all sites, on the welcome card or with the switch in Site access. Chrome asks once, and hold-key drag then works on every website. Add sites where it should never run with Never on this site; they're listed under Never on these sites.
One site at a time: Without all-sites access, turn on Hold-key drag on this site while you're on a site. Chrome asks for access to that one site.
The key: Hold a letter (Z unless you change it), or choose Command on a Mac or Ctrl on Windows and Linux. With Command or Ctrl, a plain click still opens the link as usual; only a drag starts a selection.

Typing in a text box, search field or editor never starts a selection. Pages that were already open pick up the change without a reload. If you remove Link Meteor's access in Chrome's settings, it turns hold-key drag off wherever it no longer has access. Removing all-sites access can also remove access to sites you allowed one at a time.

After a drag

Choose what happens when you let go, under After a drag in the side panel's collections and site settings.

Show the card: The card described above, with every action.
Copy right away: Copies the ticked links at once, as a text and URL table, URLs, Markdown or rich links. A small notice says what was copied; Show links opens the card.
Add right away: Adds the ticked links to the active collection at once. The notice has Undo.

The notice closes by itself after a few seconds, or with Esc. Leave out navigation links and Skip links already saved apply to these too.

Pages, tabs and windows

In the side panel, choose how much to capture, then press the capture button. The button always says exactly what it will do, such as Capture 3 tabs.

This page: Every loaded link on the page you opened Link Meteor from, including links inside components, open or closed, frames from the same site, and frames from other sites that Link Meteor has access to; see frames and closed components. Scroll to the end first and Follow Next go further than one screen; see more than one screen or page. On a PDF's tab, the button reads Capture this PDF; see the links inside PDFs. If the tab has since moved to another site, Link Meteor asks Chrome for access to it in the same click, unless you've allowed all sites; see troubleshooting.
Pick tabs: Choose exact tabs from a list of your open tabs, grouped by window.
This window: Every ordinary tab in the current window.
All windows: Every ordinary tab in every normal window of your current Chrome profile.

The first time you choose anything other than This page, Chrome asks whether Link Meteor may see your list of tabs. When you capture, Chrome also asks for access to the sites involved. If you decline, those pages are reported as denied; everything else is still captured. Tabs opened after you start a capture aren't added to it, and one capture can include up to 100 tabs.

Save the tabs themselves. Choose The tabs themselves instead of Links in the pages to save each tab as one link, with the tab's title as its anchor text, instead of the links inside it. This works with every scope, from This page to All windows. Tabs that aren't web pages, such as Chrome's settings, are skipped and counted. No site access is needed; picking tabs and whole windows need the tab list, as above.

The links inside PDFs

Link Meteor reads the links inside a PDF: each with the words under it, the page it is on and the words around it. In a reference list, the words around a link are its whole reference. You see every link before anything is saved.

From a PDF's tab: When the tab you're on shows a PDF, the capture button reads Capture this PDF. Link Meteor asks Chrome for the PDF's file again from the same address, as that page; Chrome usually answers from its own cache. The PDF is read in your browser by PDF.js, which is part of Link Meteor, and is never uploaded. The toolbar's temporary access covers it, so no new permission is needed. Regions can't be drawn on a PDF, so Select a region is off there.
From a PDF file: In Import links, choose a PDF, or drop one on the Import links panel or on the list. It is read in your browser, never uploaded, and needs no access at all. Its links are marked Imported, with the file's name and their page, such as paper.pdf, page 3.
A PDF on your computer: Chrome doesn't let an extension read a PDF you opened from your computer through its tab, so there the button reads Choose this PDF… and opens the file chooser. You can also drop the file into Link Meteor's side panel or full view. Link Meteor asks for no access, and doesn't need Chrome's Allow access to file URLs.
The preview: Lists every link with its page, its words and its address, and what the PDF says about itself: arXiv's stamp, a DOI and its title, each with where it was read. Show what was left out lists the links to places inside the PDF, addresses that aren't web, email or phone addresses, and links already saved there. Choose the collection, or a new one named after the PDF. Also save this PDF as a link adds the PDF itself, with what it says about itself; for a file, that's the paper's arXiv page or DOI address, when the PDF names one. Add N links saves one batch, and the notice has Undo.
When it can't be read: A PDF that needs a password, a damaged file and a file over the limit each say so, and nothing is added. So does a PDF with no links Link Meteor can read: scanned pages are pictures, and an address printed without a link isn't a link. If the site won't send the PDF again, such as a sign-in page or an expired address, download the PDF, then import the file. Cancel or Esc stops the reading.

PDFs of up to 50 MB and 2,000 pages are read. A link over a picture keeps empty anchor text, as on web pages.

More than one screen or page

Three captures go further than what's on the screen. Each one starts only when you click, has a limit stated before it starts, and shows a counter and Stop on the page and in Link Meteor the whole time; Esc on the page stops it too. What it found is saved as one batch with one Undo, and is kept if you stop. It ends with a report of every page: what was found, what was added and how it ended. One runs at a time.

Scroll to the end first: A choice under This page, for pages that load more as you scroll. Link Meteor scrolls the page one screen at a time, keeps every link it passes, even ones the page later removes, then scrolls back to where you were. It stops at the end, at 50 screens, at 5,000 links or after 2 minutes.
Follow Next, up to 20 pages: A choice under This page. After each page, Link Meteor moves your tab to the page's own Next link, waits for it to load, pauses 1.5 seconds and captures it, so you see every page it reads. Set any number of pages from 2 to 20. It stops when there is no Next link, when Next is a button without an address, when the next page is on a site Link Meteor can't read, at a page it already read, or at the limit. The toolbar's temporary access lasts while the tab stays on the same site, so one click on the toolbar icon is enough there. With Scroll to the end first also on, each page is scrolled first.
Capture their pages…: Select up to 20 links in the list and choose Capture their pages…, or Capture its page… in a link's details. Each page opens in a background tab, one at a time and 2 seconds apart, is captured and closed. A PDF among them is read as a PDF while Link Meteor stays open. It goes one level deep: the links it finds are saved, never followed. The panel names the sites Link Meteor can't read yet, with one Allow these N sites, and asks which collection the links go to, or a new one.

Leave out navigation links and Skip links already saved apply to every page of a run. Email and phone links, and links to files that aren't pages, are left out of Capture their pages…, and the panel says how many.

Frames and closed components

Closed components: Some sites build parts of their pages as closed components, which page scripts can't look into. Link Meteor reads the links inside them, for Capture this page, a region, hold-key drag and the right-click menus. It needs no permission for this.
Frames from other sites: Capture this page reads every frame Link Meteor has access to. The toolbar's temporary access covers the page's own site, not frames from other sites, so the report says how many frames weren't read and names their sites. Allow these sites asks Chrome for exactly those sites, and their links then join the same capture, with the same Undo. Up to 20 sites are offered at a time. With all-sites access, every frame is read.
Frames still loading: A frame that hadn't loaded yet is counted and said to be loading. Scroll to it, let it load, and capture the page again.

A region, hold-key drag and the right-click menus stay in the page itself. They read frames from the same site, and the card says how many frames from other sites a region touched but couldn't include.

Right-click menus

Link Meteor's actions are under Link Meteor in Chrome's right-click menu. What's offered depends on where you click:

On a link: Add link to “collection” saves that one link to the active collection, with its anchor text. Copy link text + URL copies it as two spreadsheet columns. Download linked file appears on links to files, such as PDFs and images. Select a region is there too.
On selected text: Capture links in the selection saves every link the selection touches.
On the page: Select a region, Capture this page and Save this tab as a link.
On the toolbar icon: Save this tab as a link, Save all tabs in this window as links and Open the full view.

A save shows a small notice on the page with Undo and Show links. Saving all tabs needs the tab list: without it, the full view opens with that choice ready and Chrome asks when you save.

On a PDF. Regions can't be drawn on a PDF, so Select a region says so and points to Capture this PDF in the side panel, which reads every link in it; see the links inside PDFs. Capture this page lists the tab as a PDF in its report. Add link still works: right-click a link in the PDF and it is saved with its address. The same goes for files on your computer.

The capture report

After a page or tab capture, a report lists every page with its own result. Choose Show only these links to filter the list to that capture. After a capture of more than one page, the report is a table of every page with what was found, what was added and how it ended, with Undo for the whole run.

Captured: The number of links saved from that page, with notes about anything skipped, such as malformed links or frames from sites Link Meteor has no access to, with Allow these sites to include them.
No links: The page loaded and has no links Link Meteor can read. This is different from being denied access.
Access denied: Link Meteor doesn't have permission for that site. Capture again to be asked, or allow access in Chrome's extension settings.
PDF: A PDF's tab in a capture of tabs. Capture it by itself with Capture this PDF.
Unsupported page: Chrome's own pages, the Chrome Web Store and incognito tabs can't be captured.
Capture failed: The tab closed, navigated to a new page mid-capture, or another error happened. The message says which.

What gets saved

Each captured link is a separate occurrence, even when the same URL appears more than once. Every occurrence keeps these fields apart:

Anchor text: The link's visible text, with runs of spaces and line breaks collapsed to single spaces and the ends trimmed. Hidden text inside a link is left out. A link with no visible text has empty anchor text.
Accessible label: The link's ARIA label, or the alt text of images inside it. Kept separate from anchor text.
URL: The full destination the link points to. Web, email (mailto:) and phone (tel:) links are captured; script links such as javascript: are not.
Original href: The link's address exactly as written in the page, which may be relative.
Source page: The address and title of the page it was captured from, plus the frame's address if the link was inside a frame. For a link read from a PDF's tab, the source page is the PDF.
PDF page: For a link read from a PDF, the page it is on, counting from 1. A link's details say “Page 3 of the PDF”, and exports have a PDF page column.
Capture time and batch: When it was captured, and an ID shared by everything captured in the same action.
Notes and tags: Yours to add. They are stored separately and never change what was captured.
Custom columns: Your own fields for a collection, such as a deadline or a contact, filled in by hand. See Collections.
Context: The words around the link on its page, up to 400 characters, kept only where they are a sentence or a paragraph. A bare list of links has none. In a PDF, it is the line or lines the link sits on, or its whole reference in a reference list. Turn this off with Save the words around each link, under After a drag.
Reading status and star: Unread, Reading or Read, and a star. Yours to set.
Imported from: For links you import, the file or bookmark folder they came from, and for a PDF file, the page, such as paper.pdf, page 3.

Link Meteor never replaces a label with a page title. Authors, dates and DOIs come from what the pages and PDFs you had open say about themselves, or from a link's own address. Only if you turn on Page details lookup and click Look up details do they also come from Crossref, DataCite or NCBI.

Context and citations

For research, Link Meteor keeps more than the link, read from the pages and PDFs you had open:

Identifiers: A DOI, arXiv ID, PubMed or PMC ID, or ISBN found in a link's address. Rows show them as badges, and a link's details list each one with Copy.
Page citations: Journal sites, arXiv, PubMed and many others publish citation tags about each page: title, authors, journal, date, pages and identifiers. Link Meteor keeps them for the page you capture from, and for tabs you save as links. Saving a tab reads its tags only where Link Meteor already has access to the site.
Borrowed citations: A link uses the citation of a saved page about the same work: a page that names the link as its PDF, or one with the same DOI or arXiv ID. So capturing View PDF on an arXiv abstract page gives the PDF link its title, every author and the date. A link's details show the citation and where it came from.
What a PDF says about itself: When you read a PDF's links, Link Meteor keeps the PDF's citation: arXiv's stamp on its first page (the ID, version, category and date), a DOI from its own details or its first page, and its title. Authors only when the PDF's own details name them and each name is printed on its first page; PDFs rarely name them reliably. A page's own citation tags win over what a PDF says; the PDF fills only what they leave out.
arXiv papers: When nothing else dates an arXiv link, its year and month come from its ID, and citation exports mark it as an arXiv preprint.

A link's details show its citation and say where it was read: the page's own tags, the PDF itself, or Crossref, DataCite or NCBI after a lookup. A link saved on its own, such as a PDF tab, may have no authors yet. Capture from the paper's own page, save that page as a tab, or use Page details lookup, and they fill in. The Export panel says how many papers still have none.

Page details lookup

Page details lookup fills in a paper's missing title, authors, date and journal from its DOI, arXiv ID or PubMed ID. It is off unless you turn it on, under Online lookups in the side panel's collections and site settings, which says exactly what it sends and where. Until you turn it on, Link Meteor requests nothing from any service.

What it sends, and where: Only the paper's DOI, arXiv ID or PubMed ID, and nothing else: DOIs to Crossref (api.crossref.org); DOIs Crossref doesn't have, including arXiv's, to DataCite (api.datacite.org); and PubMed IDs to NCBI (eutils.ncbi.nlm.nih.gov). No page address, link text, note or collection name is sent, and no cookies are sent or kept. The privacy page has the details.
Where to start it: Look up details is in the list's toolbar for selected links, in a link's details when its citation is missing something, and in the Export panel beside the line that counts papers with no authors yet. It never runs by itself. With the switch off, the button sends nothing and shows you the switch.
What it changes: Only what is missing. It never replaces what a page's own citation tags gave; details read from a PDF itself can give way to the lookup's, and the result says so. It sends one request at a time, with a counter and Stop, and up to 200 papers in one go. When it ends, it says what it filled in, and Undo puts the details back as they were.

When a citation came from a lookup, a link's details name the service and the date it was looked up. Citation exports use what the lookup filled in.

Review and curate

Search matches anchor text, URLs, source pages, context, notes, tags and custom column values. Press / to jump to the search box.
View opens filters for domain, file type (such as pdf) and whether links stay on the source site, plus sorting and grouping. Active options appear as chips you can remove one by one.
Group shows one row per unique URL, or per unique URL and anchor text. Grouping is only a view: each group lists every occurrence, label and source page, and switching back shows them all again.
Details: open any row to see every saved field, its context, identifiers and citation, and add notes, tags and custom column values to that occurrence.
Reading status and stars: in a link's details, mark it Unread, Reading or Read, and star it. Mark as read, Mark as unread and Star do the same for selected links, each with Undo. Under View, Reading and Starred only filter the list.
Insights: switch from Links to Insights to see a collection's sites, file types, reading status, repeated addresses and captures by week. Choose any entry to see those links.
Move to… and Copy to… send the selected links to another collection or a new one. Notes, tags, reading status, stars, custom column values and citations go with them, and a column the other collection lacks is created there. A link that's already there stays where it was. Undo puts both collections back. Move to… is also in each link's details.
Fill for selected links sets one custom column to the same value on every selected link, and says how many values it replaces. Undo puts the earlier values back.
Select and remove: tick rows, or tick Select page. When the view has more than one page, Select all then selects every link in it. Remove the selection, or use Remove all in this view for everything that matches your search and filters. Undo puts removed links back in their original positions.

Long collections are shown 100 rows at a time. Selections carry across pages.

Collections

Every capture goes into the active collection. Create collections for different projects, switch between them from the collections list, and use the pencil button to rename a collection or add notes and tags to it. Deleting a collection asks for confirmation and can't be undone.

Empty this collection removes every link but keeps the collection's name, notes and tags, and can be undone.

Custom columns. In the collection editor (the pencil button), add up to 20 columns of your own, each with a name of up to 60 characters. Fill them in each link's details. Rename a column there, or remove it with its values; Undo brings both back. Search, exports and backups include them.

Collections are saved on your computer by Chrome and survive restarts. They aren't synced to other computers. Removing Link Meteor from Chrome deletes everything it saved. A backup file keeps it:

Download a backup: Saves one file with every collection, link, note, tag and setting to your downloads folder. Backup and restore shows when you last made one.
Restore from a backup: Choose a backup file, on this computer or another one. A preview shows what will change before anything does. Merge adds the backup's collections and links to yours, skipping links already saved, and joins custom columns that share a name. Replace swaps everything for the backup. Either way, Undo puts things back as they were. Version 1.0.0 restores backups from 0.3.0 on; a backup made with 1.0.0 can't be restored by 0.5.0 or earlier.

Import links

Import links, in the side panel's collections and site settings, brings links in from:

a file of up to 20 MB: CSV, TSV, Excel, a text or Markdown list, an HTML page or a browser's bookmarks file, or a Link Meteor JSON export;
a PDF of up to 50 MB and 2,000 pages: the links inside it, each with its page (see the links inside PDFs);
text you paste: addresses, Markdown links or HTML;
a Chrome bookmark folder, with or without its subfolders. Chrome asks for bookmark access the first time.

Choose which column holds the addresses and, if you like, which hold the anchor text, notes, tags, reading status, star and custom columns. A preview lists what will be added and every skipped row with its reason. Up to 20,000 links import at once, into a collection you have or a new one, and Undo removes them. Choose a file, or drop one on the Import links panel or on the list. Files are read in your browser and never uploaded.

Copy and export

What gets exported: every link in the current view, across all pages, unless you've selected some, in which case only the selected links that still match your filters. The export panel always shows the exact count.

Columns: anchor text and URL by default. Add any saved field or custom column and change the order; the table copy and the CSV, TSV, Excel and HTML files follow it. New custom column… at the end of Add a column creates one.

File names: the collection's name with the date and time, such as Urban-heat-islands-sources_2026-09-26_1432.xlsx. Set an optional prefix, or the date only, under File name pattern. Type in File name to name one download.

Copy table, URLs or Markdown: Tab-separated columns, one URL per line, or one [anchor text](URL) per line.
Rich links: Links that keep their anchor text clickable when pasted into Google Docs, Word or Notion, with plain text for everywhere else.
Excel (.xlsx): A real workbook with a bold header row that stays in view, filter buttons and clickable links. An About sheet records when it was exported, from which collection, and with which view and filters. Every cell is text, so nothing is turned into a formula, number or date.
CSV and TSV: Properly quoted. Cells that look like formulas (starting with =, +, - or @) get a leading apostrophe so spreadsheets import them as text.
Markdown and HTML: Page text is escaped so it displays as text and can't act as markup.
JSON: Every field for every link, with an about block like the Excel About sheet. In grouped views, each row includes all of its occurrences.
URL list: One URL per line, exactly as captured.

In a grouped view, tables, text and Markdown use each group's first link; JSON keeps every occurrence and source.

Columns you can add include context, reading status, starred, DOI, arXiv ID, PubMed ID, ISBN, imported from and PDF page.

Citations and notes

The Format list also writes one entry per link for reference managers and notes:

BibTeX (.bib): For BibTeX, biblatex, Zotero and BibDesk. Acronyms and words with inner capitals are braced, so styles keep them as written.
RIS (.ris): For Zotero, EndNote and Mendeley.
CSL-JSON (.json): For Zotero, Pandoc and other citation tools.
Annotated bibliography (.md): A numbered list: each citation with your note, its context as a quote, and its tags.
Obsidian note (.md): A note with properties, one list item per link, your tags, custom columns as name:: value fields, and context quotes.

Titles, authors and dates come from each link's citation (see Context and citations); without one, the title is the anchor text. The panel shows the first entry and says how many entries have a DOI or arXiv ID, how many have authors and a date, and how many papers have no authors yet, with Look up details beside it (see Page details lookup). Exporting never looks anything up online.

Bookmarks and opening links

Save as bookmarks creates a new folder in Chrome's bookmarks, or adds to an Existing folder you find by name. Skip links already in that folder is on by default. Each bookmark is named with the anchor text, or the URL if a link has no text. Chrome asks for bookmark access the first time. Only web links are bookmarked, and Chrome may sync bookmarks to your other devices according to your own browser settings.

Open in tabs opens web links as background tabs; you can also open them in a new window or as a tab group named after the collection. Up to 500 links open at once. Above 20, Link Meteor asks first, and above 100 it suggests a new window or a tab group. Chrome asks once for tab-group access, to name the group. Opening a link visits that site, like clicking it.

Download the files behind links

Save the files that links point to, such as PDFs, images, documents and data files, as if you had clicked each one:

on the card, More, then Download N files (F), for the ticked links;
in the Export panel, Download files, for the links in the view or your selection;
in a link's details, Download, for that one link;
by right-clicking a file link, Link Meteor, then Download linked file.

A link counts as a file when its address ends in a file type, such as .pdf or .png, when its text starts with [PDF], or when it's a known PDF address, such as arXiv's. Files are saved in Downloads/Link Meteor/, in a folder named after the collection. Each is named after its anchor text, or after its address when the text is only a generic word such as “pdf”, “Download PDF” or “Full text”, so arXiv's PDFs save as 2409.11211v1.pdf, not pdf.pdf. Every name fits in 255 bytes. A PDF linked at several of its pages, such as #page=3 and #page=9, downloads once.

The first time, Chrome asks for download access, in the full view; the card and the menu send you there. Link Meteor downloads three files at a time and asks first above 10. Cancel stops what hasn't finished. When a link gives a web page instead of a file, often a sign-in page for a paywalled article, the result says so. A download that fails says why.

Downloading visits the file's address with your browser's usual cookies, like clicking the link, and only when you ask. Link Meteor never reads, opens, changes or removes your other downloads.

PDFs as one ZIP or one PDF

In the Export panel's Download files, the PDFs among the links in the view or your selection get two more actions:

Download as one ZIP: Saves the PDFs in one ZIP, each file exactly as the site sent it and named as Download files would name it. The ZIP is named like an export, such as Thesis_2026-10-02_1412.zip.
Combine into one PDF…: Lists the PDFs in the view's order. Move each one up or down, or leave it out, and add PDF files from your computer with Add PDF files… or by dropping them on the list. Combine writes one PDF that keeps every page, its text, its links and each part's bookmarks. Page labels (such as i, ii, iii), form fields and attachments may not carry over. A PDF that needs a password can't be combined, but it can go in a ZIP as it is.

Link Meteor gets each PDF through a page of its own site, as when you click the link, signed in as you are: one at a time, 2 seconds apart, with a counter and Stop. PDFs on the site of the tab you're on need nothing more, once Link Meteor can read that tab. For other sites, unless you allowed all sites, Chrome asks for exactly those sites when you start, and a PDF on a site you don't allow is left out. The files are put together in your browser and saved through Chrome's download, like an export; nothing is uploaded, and Link Meteor keeps no copy. Keep Link Meteor open until it's done. The result lists every PDF with how it ended.

Up to 20 PDFs and 200 MB at a time, 50 MB each, and a combined PDF of up to 5,000 pages. Download files itself still saves each file on its own.

Appearance

Under Appearance in the side panel's collections and site settings, choose one of seven themes: Meteor, Comet, Aurora, Ember, Nebula, Graphite or High contrast. Then choose System, Light or Dark; System follows your computer's setting. The theme also colors the capture card and highlights on web pages, and the toolbar icon. This website has the same themes in its theme menu.

Limits and coverage

A capture reads the links loaded in the page. Scroll to the end first, Follow Next and Capture their pages… go further only when you start them, within their limits: 50 screens, 5,000 links or 2 minutes for a scroll, and 20 pages for the other two. Link Meteor never crawls: the links it finds are saved, not followed.
Frames from other sites are read only where Link Meteor has access to their sites, and the report names the ones it couldn't read. A region, hold-key drag and the right-click menus don't reach into them. Buttons that navigate with scripts but aren't real links aren't captured.
A PDF can be read when it is up to 50 MB and 2,000 pages and needs no password. Only real links in it are read: scanned pages are pictures, and an address printed without a link isn't picked up. A PDF on your computer is read from its file, not from its tab.
Chrome's own pages, the Chrome Web Store and incognito windows are off limits to extensions.
One capture can include up to 20,000 links from a page or region and up to 100 tabs. Larger captures are kept up to the limit and marked as partial.
One opening action can open up to 500 links, and one download action can save up to 100 files. One ZIP or combined PDF holds up to 20 PDFs and 200 MB, and a combined PDF up to 5,000 pages.
One import can add up to 20,000 links, from a file of up to 20 MB or a PDF of up to 50 MB.
Page details lookup asks about up to 200 papers at a time, one request at a time.
Collections use Chrome's local storage for extensions, without its usual 10 MB limit. If a save ever fails because space runs out, your existing collections are left intact and Link Meteor says so. A backup larger than 50 MB can't be restored in one step.

These are practical limits of the extension, not paid tiers. Every feature is free.

Keyboard

⌥ ⇧ L / Alt Shift L: Start selecting a region on the current page. Change it at chrome://extensions/shortcuts.
Esc: Cancel a selection, close a confirmation, or go back from a panel view. On a page Link Meteor is scrolling or capturing, stop the run. While a PDF is being read, stop the reading.
/: Jump to the search box.
Tab: Every control can be reached and used with the keyboard, with a visible focus outline.