Collections Plus

An open, local-first replacement for Microsoft Edge Collections, which is being retired in Edge 149 (~June 2026). 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.

What it does

Saving

• Side panel that mirrors the original Collections pane.
• 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.
• Auto-file rules: route quick-saves (Ctrl+Shift+S) into a collection automatically by domain, URL or title (⋯ → Auto-file rules…).
• 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).
• Sort & compact view: sort items (Newest / Oldest / A–Z) and collections (Newest / A–Z), or keep your manual drag order; a compact toggle fits more 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.
• 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 (with a toolbar count); opening one marks it read, or use Mark all read. Bulk imports stay read so they don't flood it.
• Light / dark / system theme: cycle in the ⋯ menu. System follows your OS / Chrome light-dark setting and switches live (e.g. at sunrise/sunset).
• Command palette (Ctrl+K / ⌘K): jump to any collection or run any command from one keyboard-driven, fuzzy-filtered list.
• Address-bar search: type col + space in the browser's address bar to search your saved pages and jump straight to one.

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.
• 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 (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.

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

An unpacked extension does not auto-update; each browser keeps running the code it last loaded until you hit reload ↻ on its card. So 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. (Publishing to the Chrome Web Store / Edge Add-ons would enable background auto-updates; until then, reloading per device is just how unpacked extensions work.)

Migrating from Edge Collections

Before Collections is removed, open the Collections pane in Edge and click Export Your Data. Edge writes collections_export.csv to your Documents folder. Then in this extension:

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

The CSV only contains saved pages (Edge's export drops images and notes), so the import is pages-only. Going forward, use 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.

Reorder collections and items

Drag the ⠿ handle on a collection card to reorder your list, and the handle on an item to reorder it within a collection. The new order is saved instantly.

Each card's actions (▶ open, 📁 move to folder, 📦 archive, 📍 pin, 🗑 trash) stay hidden until you hover the card, so long collection names get the full width and wrap in full instead of being clipped. Pinned collections show a small 📌 badge at rest.

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 reachable from the 📦 Archive and 🗑 Trash icons at the top of the panel (each shows a small count badge when it has something in it).

• Trash is for things you've deleted. The 🗑 button on a collection card — or ⋯ → Move to Trash inside a collection, or 🗑 on a folder header — 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. The 📦 button on a card (or ⋯ → Archive collection) moves a collection to the Archive; nothing there is ever auto-deleted. Restore brings it back to the top of your list.
• 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 (CSV)

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

The file opens straight in Excel / Google Sheets / Numbers: one row per item, with columns Collection, Type, Title, URL, Note, Added. It's a UTF-8 CSV (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 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:

⋯ → Chat with collections (AI)… for all collections, or inside a collection ⋯ → Chat about this collection (AI)… 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.

Syncing across devices

Sync is optional and provider-agnostic: there's no account and no API keys. Instead of integrating with one cloud (OneDrive vs. Google Drive vs. …), 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.

On your first device:

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

On every other device:

⋯ (top-right) → 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.

How it stays in sync

Two simple operations run against that one file:

• Push: when you change anything, a moment later the extension writes your collections to the file. Your cloud client carries the file to your other devices.
• Pull: the extension reads the file back and adopts it if it 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: the timestamp your own filesystem records when the cloud client drops in a new copy. Crucially, this does not depend on your two computers' clocks agreeing, so a change made on one machine can't be mistaken for "older" and skipped on another. (Earlier builds compared a clock value embedded by the writing device, which could silently drop changes when clocks drifted; that's fixed.)

Conflicts & safety

Reconciliation is last-write-wins: whichever device writes the file last is the version everyone converges on. That's exactly right for one person moving between a laptop and a desktop. It is not built for two people (or two offline devices) editing the same collection at once: if you edit on two devices while both are offline, the one whose file syncs last wins and the other's un-synced edits are overwritten. So let the cloud catch up before editing elsewhere, and keep Export backup (JSON) as a guaranteed-complete snapshot.

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, so you can't accidentally clobber the good copy. And if a device has un-pushed local edits when the file changes elsewhere, it won't silently overwrite them; it keeps your changes and offers "Use file instead." The sync menu shows when you last synced, and a local version history (⋯ → Version history…) lets you roll back recent states.

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 ⋯ → 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.

Data & privacy

Everything is stored locally via chrome.storage.local on your machine. Nothing is sent to us or any server; the optional sync file lives in a folder you chose, and only your own cloud client 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. The AI chat is the only feature 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: side panel, context menus, shortcut
sidepanel/           panel.html / panel.css / panel.js (the UI)
lib/store.js         data layer (chrome.storage.local), schema, settings, history
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 (pure parts testable)
lib/xlsx.js          dependency-free .xlsx writer (pure, testable)
lib/image.js         downscale a file/URL/screenshot into a small data URL
lib/sync.js          optional synced-file sync (File System Access API)
icons/               generated PNG icons
tools/make_icons.py  regenerate the icons
tools/test_*.mjs     pure-logic test harnesses (no browser needed)
fixtures/            sample Edge CSV for testing the importer

Development

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

  npm test            # csv import, csv/xlsx export, render, store/migrate
  

• Regenerate icons:

  python tools/make_icons.py
  

• 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).