Collections Plus

An open, local-first replacement for Microsoft Edge Collections, which Microsoft has now retired. Collections Plus is the way to get that workflow back — and, if you never exported before the feature disappeared, the way to recover the collections Edge left behind on disk. It's a small Manifest V3 browser extension for Chromium-based browsers: no account, no server, no build step. Your data stays in your browser. A companion mobile web app at getcollectionsplus.com/app puts the same collections on your phone, synced through your own Dropbox.

📖 Guides: getcollectionsplus.com/help

What it does

Saving

• Side panel that mirrors the original Collections pane — or a floating pop-up window if you prefer (⋯ → Tools → Open in). Esc closes either one.
• Save the current page (with thumbnail, favicon and title) with one click, or with the Ctrl+Shift+S keyboard shortcut (rebindable).
• Right-click any page, link, or image → Save to Collections Plus ▸ pick a collection (or create a new one on the spot). Select text → Save selection as note, or Save selection as highlight to keep the quote with a link back to its source and your own annotation.
• Add all open tabs in the window to a collection at once (inside a collection, ⋯ → Add → Add all open tabs).
• Auto-file rules: route quick-saves (Ctrl+Shift+S) into a collection automatically by domain, URL or title (⋯ → Tools → Auto-file rules…). They also apply to links shared to the mobile app.
• Drag a link or image straight onto the panel to save it.
• Duplicate-aware: saving a page already in the collection is skipped.
• When a page has no preview image, a local screenshot is captured as the thumbnail.

Organizing

• Open a whole collection in one click: the ▶ button on a collection card opens every saved page at once, gathered into a named browser tab group (the collection's name — the old Edge Collections behavior). Also available via ⋯ → Open all pages inside a collection.
• Free-form notes; drag-and-drop reorder of both items and collections.
• Search across every collection (titles, URLs, notes, tags, and custom fields); results show which items matched, and an open collection gets its own filter box to narrow its items instantly.
• Folders to group collections, pin favorites to the top, and tags (click a tag to filter). Reorder folders by their drag handle, and — in manual sort — folders and top-level collections share one order, so you can interleave them freely (Folder A, Collection 1, Folder B, …).
• Sort & compact view: sort items (Newest / Oldest / A–Z) and collections (Newest / A–Z), or keep your manual drag order; a ▥ compact toggle on each list (items and collections) fits roughly twice as many on screen. Sorting is display-only and your choice is remembered.
• Checkboxes turn any collection into a checklist; custom fields (price, qty, SKU…) add structured data that flows into exports. Field values are multi-line, so one doubles nicely as a per-item comment box.
• Move or copy items between collections; undo deletes.
• Trash & Archive: deleting a collection or folder moves it to the Trash (restorable, and auto-emptied after 30 days), and you can Archive old collections to keep your main list uncluttered. Both are in the ⋯ menu.
• Custom covers: upload an image or promote a saved thumbnail with ★.
• Reading list (read-it-later): saved pages start unread and gather in the 📖 Reading list (in the ⋯ menu, with its count mirrored on the ⋯ button); opening one marks it read wherever you open it from, or use Mark all read. Bulk imports stay read so they don't flood it. Don't want the mark-on-open? Mark read when opened: Off (⋯ → Tools). Don't use the feature at all? Reading list: Off turns it off entirely.
• Light / dark / system theme: pick one under Theme in the ⋯ menu. follows your OS / Chrome light-dark setting and switches live (e.g. at sunrise/sunset).

Exporting & backup

• Export to Excel as a real .xlsx workbook (one sheet per collection, clickable links) or as CSV, great for part lists, shopping lists, or anything you want to sort and total.
• Export to Markdown or HTML, or copy a collection's links to the clipboard.
• Share as a web page: turn any collection into a clean, self-contained HTML page (responsive cards with thumbnails, links, notes and tags) you can send to anyone or host anywhere — saved as one portable file, with a one-click preview.
• Import your Edge export (collections_export.csv) so your existing collections carry over on day one. A CSV has no pictures, so imported pages come in blank — click Fetch missing images (in a collection's ⋯ menu, or Fetch all missing images in settings) to pull each page's preview image in bulk. Or hover a single page and click 🖼️ to fetch just that one. By default it only fills in pages with no image yet.
• Import Edge's leftover database (for anyone who missed the CSV export before Collections was retired). Edge leaves a SQLite file behind at %LOCALAPPDATA%\Microsoft\Edge\User Data\<profile>\Collections\collectionsSQLite — ⋯ → Import Edge database (SQLite)…, browse to that file, and your collections come back, with thumbnails and favicons (which the CSV never had). It's read directly with a small built-in SQLite reader — no extra tools.
• Import your browser bookmarks (Chrome/Edge) — each bookmark folder becomes a collection.
• JSON backup (high fidelity: keeps notes, images, fields) and a local version history you can roll back to.
• Optional offline image caching so saved images survive link rot.
• Dead-link checking & content snapshots: Check links (per collection or all at once) probes your saved pages and badges any that have gone ⚠ dead; turn on Auto-check links to have it run quietly in the background. The 📄 snapshot button on a page saves a readable copy of the article text so the content survives even if the original goes offline — read it back anytime.
• Optional cross-device sync through a single file you keep in any cloud-synced folder, merged item by item so edits on different devices never overwrite each other (see Syncing across devices).

Ask your collections (optional AI)

• Chat with your saved data. Connect your own generative-AI key and ask for summaries, reports, comparisons, or "what did I save about X?" — answered from the collections you saved.
• Bring any provider: Claude (Anthropic), OpenAI, Grok (xAI), Gemini (Google), Azure AI Foundry, or a local Ollama model. Scope the chat to a single collection or all of them.
• Put the AI to work, not just to chat: Summarize (AI) a saved page snapshot into a TL;DR note, Suggest tags (AI) for a collection (you approve what's added), or Organize this collection (AI) for grouping, duplicate and title suggestions.
• AI search (✨): find items by meaning across everything you've saved — not just keyword matches — with a one-line reason and a jump to each item's collection. The instant keyword search stays available alongside it.
• Weekly digest (AI): a friendly recap of everything saved in the last 7 days, grouped by collection, rendered in the chat so you can ask follow-ups.
• Your key stays local (chrome.storage.local), is never synced or exported, and the whole feature is off until you set it up. See Ask your collections and the Privacy Policy.

Browser support

This is a standard Chromium Manifest V3 extension, so the core (context-menu saving, storage, page capture, import/export) runs on any modern Chromium browser. The catch is the Side Panel API (chrome.sidePanel), which hosts the whole UI; its support varies:

Browser	Side panel UI	Notes
Google Chrome (114+)	✅ Full	Verified.
Microsoft Edge (114+)	✅ Full	Verified.
Brave	⚠️ Partial	Panel opens but a known bug can dismiss it after ~1s.
Vivaldi	⚠️ Broken	Ships its own "Web Panels"; the standard API doesn't work.
Opera	❌ No	Doesn't implement the extension Side Panel API.
Arc / other Chromium	➖ Likely	Should work if built on Chromium 114+ with the Side Panel API.

In short: Chrome and Edge are the recommended, fully-tested targets. Other Chromium browsers may work depending on their Side Panel API support.

Install

From the Chrome Web Store

Collections Plus is live on the Chrome Web Store. Install it in one click and it auto-updates, with no manual reloading.

🛈 Install Collections Plus from the Chrome Web Store →

Load unpacked (available now)

1. Clone or download the repo from https://github.com/rod-trent/Collections-Plus.
2. Open edge://extensions (or chrome://extensions).
3. Turn on Developer mode (toggle, usually bottom-left or top-right).
4. Click Load unpacked and select the project folder you just cloned.
5. Pin the Collections Plus icon to the toolbar and click it to open the panel.

To update after pulling changes, click the reload ↻ button on the extension's card in edge://extensions.

On your phone

Open getcollectionsplus.com/app and add it to your home screen (Android: Chrome's Install app; iPhone: Safari's Share → Add to Home Screen). See Mobile app below.

Keyboard shortcut

Ctrl+Shift+S saves the current page to the active collection. To change it, open edge://extensions/shortcuts (or chrome://extensions/shortcuts) and rebind "Save the current page to Collections Plus."

Updating across devices

The Chrome Web Store version updates itself in the background. An unpacked extension does not: each browser keeps running the code it last loaded until you hit reload ↻ on its card. So if you run it unpacked, whenever you pull new code (or apply a fix), reload the extension on every device, not just one. This matters especially for sync: a device still running older code can not only miss changes but also push with the old logic, so reload all of your devices around the same time after an update. Reloading only reloads the code: your collections and the sync file are untouched. The mobile app updates itself the next time it's opened.

Migrating from Edge Collections

Edge Collections has been retired, so the export button is no longer there to click. Which path you take depends on whether you got your data out in time.

If you never exported — which is most people — your collections are very likely still on disk. Edge leaves its database behind when the feature goes:

%LOCALAPPDATA%\Microsoft\Edge\User Data\<profile>\Collections\collectionsSQLite

⋯ (top-right) → Import → Import Edge database (SQLite)… → choose collectionsSQLite

That rebuilds your collections with each page's thumbnail and favicon, which the CSV never carried. The file has no extension, so browse straight to it (set the file picker to "All files" if it's hidden). This works as long as the file is still there — if you've since cleared Edge's data or deleted the profile, there may be nothing left to read.

If you did export while Collections still existed, you have collections_export.csv in your Documents folder:

⋯ (top-right) → Import → Import Edge CSV… → choose collections_export.csv

The CSV only contains saved pages (Edge's export drops images and notes), so that import is pages-only — prefer the SQLite route above if you still have the database. Going forward, use ⋯ → Export → Export backup (JSON) for a complete backup.

Organizing & exporting

Open all pages (tab groups)

A collection exists to be opened, so opening one is a single click. Each collection card has a ▶ button (it appears when you hover the card) that opens every saved page in that collection at once. The same action lives inside a collection at ⋯ → Open all pages.

The opened pages are gathered into a browser tab group named after the collection — the old Edge Collections behavior — so a whole project opens together and stays collapsible and out of the way. If you open a large collection (more than 8 pages), it asks for confirmation first. Tab grouping needs a Chromium browser with the Tab Groups API (Chrome/Edge have it); where it's unavailable, the pages still open as ordinary tabs.

Prefer the panel to get out of the way once you've opened a collection? Turn on ⋯ → Tools → Close panel after Open all. It's off by default; when on, the side panel closes itself as soon as the tabs open, so opening a collection is a single click with nothing left to dismiss.

Side panel or floating pop-up

⋯ → Tools → Open in switches where the toolbar icon opens the UI:

• Side panel (default) — docked to the browser window, as before.
• Pop-up window — the same UI in its own small window. It doesn't slide in and out or reflow the page underneath, which is the friendlier behavior in fullscreen, and it remembers where you last put it and how big you made it.

Either way, Esc closes the panel: the first press backs out of whatever's open (a menu, a dialog, a search field you're typing in), and the next one closes the panel itself, so you never have to reach for the X.

Reorder collections and items

Day to day, hovering a collection shows just the two things you do most: ▶ Open all and ＋ Add current page. To rearrange, click ⇅ next to the search box to switch to organize mode: every card shows its ⠿ drag handle, and hovering a card shows the managing buttons (📁 move to a folder or collection, 📦 archive, 📍 pin, 🗑 move to Trash). Folder headers show their color, rename and delete buttons. Click Done (or ⇅ again) to go back. The panel always opens in the everyday mode.

Drag a card's handle to reorder your list, and an item's handle to reorder it within a collection. The new order is saved instantly. Pinned collections show a small 📌 badge.

Custom covers

Every collection has a cover thumbnail (the first saved item by default). To change it:

• Open a collection and click Change cover… to upload your own image; it's downscaled and stored inline, so it always renders even if the original page goes offline.
• Or hover any saved page/image and click its ★ to promote that thumbnail to the cover.
• ⋯ → Remove cover clears it back to the default.

Trash & Archive

Two holding areas keep your main list tidy without losing anything. Both are in the ⋯ menu (📦 Archive and 🗑 Trash, each with a count when it has something in it).

• Trash is for things you've deleted. 🗑 on a collection card in organize mode, ⋯ → Move to Trash inside a collection, or deleting a folder moves it to the Trash instead of deleting it outright, with an Undo toast. From the Trash view you can Restore any entry, delete one permanently, or Empty Trash. Anything left in the Trash is permanently removed 30 days after it was deleted.
• Archive is for collections you want out of the way but intact. 📦 on a card in organize mode (or ⋯ → Archive collection inside a collection) moves a collection to the Archive; nothing there is ever auto-deleted. Restore brings it back.
• Deleting a folder sends it to the Trash too, and restoring it re-adopts its original collections (which fall back to the top level while it's trashed).
• Both Trash and Archive sync across devices and are included in JSON backups, just like your collections.

Export to Excel and CSV

• ⋯ (top-right) → Export → Export to Excel (.xlsx) or Export as CSV exports every collection.
• Inside a collection, ⋯ → Export & share → Export to Excel (.xlsx) or Export as CSV exports just that one.

The .xlsx workbook has one sheet per collection with clickable links. The CSV opens straight in Excel / Google Sheets / Numbers: one row per item, with columns Collection, Type, Title, URL, Note, Added. It's UTF-8 (with a BOM) so accented characters and emoji survive. Handy for part lists, shopping lists, research sources, or anything you want to sort, filter, or total in a spreadsheet.

Ask your collections (AI)

Collections Plus can chat with your saved data using a generative-AI provider you bring. It's optional, off until you set it up, and uses your own API key — there's no shared service and no developer-run backend.

Set it up:

⋯ (top-right) → AI → AI settings… → pick a provider, paste your API key, and (optionally) adjust the base URL or model → Save. Use Test connection to confirm it works.

Supported providers (pick whichever you already have):

Provider	Notes
Claude (Anthropic)	Defaults to a current Claude model.
OpenAI	Standard chat/completions API.
Grok (xAI)	OpenAI-compatible.
Gemini (Google)	Google Generative Language API.
Azure AI Foundry	Set your resource base URL; uses an api-key header.
Ollama (local)	Runs against http://localhost:11434 — no key needed.
Other (OpenAI-compatible)	Any endpoint that speaks the OpenAI chat API.

Chat:

⋯ → AI → Chat with collections… for all collections, or inside a collection ⋯ → AI → Chat about this collection… to scope the context to just that one.

When you send a message, the extension builds a snapshot of the in-scope collection(s) — titles, URLs, notes, tags and custom fields — and sends it, along with your message, straight from your browser to the provider you chose. Ask for summaries, reports, comparisons, or to find what you saved. Replies are rendered as formatted text.

What stays private: your API key is stored locally and is never synced and never written into exports or backups. The developer never sees your key, your messages, or your collections. The data you send goes to your selected AI provider and is handled under that provider's privacy policy and terms — review them before enabling this. Nothing is sent anywhere until you actively send a chat message. Full details in the Privacy Policy.

On your phone: the mobile app has AI search too (⋯ → AI settings to add a key, then Search by meaning in search), with Claude, OpenAI, Grok, Gemini or another OpenAI-compatible service. Keys are per device and never synced.

Syncing across devices

Sync is optional and provider-agnostic: there's no account. Instead of integrating with one cloud, the extension reads and writes a single collections-sync.json file that you place inside a folder your computer already keeps synced: OneDrive, Google Drive, Dropbox, iCloud Drive, etc. Your cloud client does the actual syncing; the extension just keeps that one file up to date. The mobile app reads and writes the same file through Dropbox.

📖 Step-by-step setup, including your phone: getcollectionsplus.com/help/sync

On your first device:

⋯ (top-right) → Sync → Create sync file… → save collections-sync.json inside your synced folder.

On every other device:

⋯ (top-right) → Sync → Use existing sync file… → open the same collections-sync.json once your cloud has finished downloading it.

The second step opens a normal Open dialog (not a save/overwrite prompt), the device adopts the synced data, and it asks for permission to write so its own future edits sync too.

With your phone: connect the mobile app to Dropbox first. It creates Dropbox/Apps/Collections Plus/collections-sync.json, and that's the file every computer should use (via Use existing sync file…). The app can only see that folder, so a sync file anywhere else won't reach your phone.

How it stays in sync

• Push: when you change anything, a moment later the extension writes your collections to the file. If another device changed the file since this one last synced, its changes are merged in first, so nothing is overwritten.
• Pull: the extension reads the file back and merges in what changed. An open panel checks on focus, when it becomes visible again, and every ~20 seconds, so changes from another device show up on their own. Sync now and Pull from sync file force a push / pull immediately.

"Did the file change?" is decided by the file's modification time as this device sees it locally, so it doesn't depend on your computers' clocks agreeing.

Merging (2.10+)

Since 2.10, sync merges item by item instead of keeping whichever copy was saved last:

• Pages, notes and collections added on any device are all kept.
• If the same item was edited on two devices, the most recent edit wins.
• An item moved to another collection lands once, not twice.
• Deletes sync too: each delete leaves a small marker (kept for 90 days) so the item doesn't come back from another device.
• Archive, Trash, folders, tags, checkmarks, custom fields and auto-file rules all sync.

Every device needs 2.10 or later. Files written by older versions carry no delete markers, so they still reconcile the old way (last write wins); if one arrives while you have unsaved edits, the extension keeps your changes and offers "Use file instead."

When you connect a device to a file that already has data, it asks before doing anything destructive: load the file (replace this device's data) or overwrite the file with this device's data. Keep Export backup (JSON) as a guaranteed-complete snapshot, and version history (⋯ → Tools → Version history…) lets you roll back recent states on one device.

Resuming after a browser restart

Browsers don't keep the file's write permission across a restart or an extension reload (a security rule of the File System Access API). When that happens, Collections Plus pauses background saving instead of erroring: the sync menu shows Sync paused, and a toast offers Resume. Click Resume (or ⋯ → Sync → Resume sync / Sync now) once, approve the prompt, and your pending changes are written immediately and sync continues. Your data is never lost in the meantime; it's just held locally until you resume.

Requires a Chromium browser with the File System Access API (Chrome/Edge have it). If it's unavailable, the sync menu says so and the rest of the extension works unchanged.

Mobile app

getcollectionsplus.com/app is a web app you install from your phone's browser, with no app store. It runs the same data layer as the extension (lib/store.js on IndexedDB) and syncs with it through your own Dropbox.

• Browse, search and edit your collections, folders, Reading list, Archive and Trash; Read through a collection a page at a time (the phone's take on Open all); checklists, custom fields, tags, sorting and drag to reorder.
• Share to save on Android: share a page, text, or a selected quote from any app to Collections Plus.
• Works offline and updates itself.

Guide: getcollectionsplus.com/help/mobile.

Data & privacy

Everything is stored locally via chrome.storage.local on your machine (and, in the mobile app, in your phone's browser storage). Nothing is sent to us or any server; the optional sync file lives in a folder you chose (or your own Dropbox, for the phone), and only your own cloud touches it. Saved page/image thumbnails are stored as references to their original URLs (not copied), so they display as long as the source stays online; uploaded covers, captured screenshots, and cached images are downscaled and stored inline so they always render. Offline image caching, cross-device sync, and the AI chat are all off until you turn them on. Apart from sync to your own cloud, the AI features are the only ones that can send your collection contents off-device, and only to the provider whose key you entered — your key is stored locally and never synced or exported. See the Privacy Policy for the full breakdown.

Project layout

manifest.json          MV3 manifest (+ keyboard command)
background.js          service worker: panel/pop-up opening, context menus, shortcut
sidepanel/             panel.html / panel.css / panel.js (the UI)
lib/store.js           data layer, schema + migrations, settings, history
lib/backend.js         storage seam: chrome.storage.local (extension) or injected
lib/merge.js           item-level sync merge with tombstones (pure, testable)
lib/sync.js            extension file sync (File System Access API) using merge.js
lib/csv.js             tolerant CSV parser + Edge-export mapper (pure, testable)
lib/export.js          collections → CSV + .xlsx sheet structures (pure, testable)
lib/render.js          collections → Markdown / HTML / link list (pure, testable)
lib/ai.js              generative-AI provider adapters + context builder
lib/semantic.js        AI search catalog + result parsing (pure, testable)
lib/xlsx.js            dependency-free .xlsx writer (pure, testable)
lib/image.js           downscale a file/URL/screenshot into a small data URL
lib/idb-backend.js     IndexedDB storage backend (mobile app)
lib/remote-adapter.js  the sync transport contract
lib/sync-engine.js     pull → merge → push over any adapter (mobile app)
lib/dropbox-adapter.js Dropbox transport: OAuth PKCE, rev-checked uploads
pwa/                   the mobile web app (served at getcollectionsplus.com/app)
site/                  getcollectionsplus.com (Next.js on Vercel), incl. /help
icons/                 generated PNG icons
tools/make_icons.py    regenerate the extension + app icons
tools/test_*.mjs       pure-logic test harnesses (no browser needed)
fixtures/              sample Edge CSV / SQLite for testing the importers

The store package (npm run package) leaves out the mobile-only modules.

Development

• No build step: edit files and hit reload on the extension card.
• Run the pure-logic tests without a browser:

  npm test            # store/migrate, merge, sync, Dropbox, import/export, AI…
  

• Regenerate icons:

  python tools/make_icons.py
  

• Run the website (and the mobile app at /app/) locally:

  cd site && npm install && npm run dev
  

• Build a Chrome Web Store / Edge Add-ons upload ZIP (runtime files only):

  npm run package     # → dist/collections-plus-<version>.zip
  

License

MIT © Rod Trent. Use it, fork it, improve it.

Roadmap ideas

• Bulk multi-select of items (move/copy/delete several at once).
• Full-text search ranking.
• Streaming AI replies and saving a chat result straight into a collection.
• An Edge Add-ons listing (already live on the Chrome Web Store).