AI & LLM Tools

AnythingLLM Browser Extension Setup Guide

J
James Eriksson
··11 min read
Install the AnythingLLM browser extension in minutes. Generate API keys, connect to your instance, and save web pages to your workspace. Works on Chrome, Firefox, Edge, Brave. Complete setup guide.
TL;DR
  • Install from Chrome Web Store or Firefox Add-ons in 90 seconds
  • Generate an API key in AnythingLLM Settings > Tools > Browser Extension, then paste your connection string into the extension
  • Right-click any webpage to save text snippets or embed full pages into your workspace
  • Works with Chrome, Edge, Brave, and Firefox; requires AnythingLLM v1.6.5 or later
  • Send password-protected and VPN-locked content directly to your instance--no third-party exposure

The AnythingLLM browser extension lets you save web content directly to your AnythingLLM workspace with a right-click. Install it in Chrome, Edge, Brave, or Firefox, generate an API key in your AnythingLLM settings, and you're capturing pages and text snippets in minutes. This is how you build a live knowledge base from the web without exposing content to third parties.

What Is the AnythingLLM Browser Extension?

The extension is a lightweight add-on published by Mintplex Labs that bridges your browser and your AnythingLLM workspace. It saves web content--text snippets, full pages, even password-protected content--and sends it directly to your instance. The extension has 10,000+ active users and a 4.4 out of 5 star rating on the Chrome Web Store, with an 81.12 KiB footprint that adds almost no overhead to your browser.

Why this matters: You're running AnythingLLM for retrieval-augmented generation (RAG). Your team needs fresh, current information from the web in your knowledge base. Manual copy-paste or file uploads kill productivity. The extension removes that friction. You find a page, right-click, choose your workspace, and the content is indexed within seconds.

Key features:

  • Save text snippets from any webpage as searchable content
  • Embed entire pages so AnythingLLM indexes every element
  • Send gated, password-protected, and VPN-locked pages to your instance without exposing them
  • Choose which workspace receives the content (unlike generic web clippers that just dump everything into one folder)
  • Works offline if AnythingLLM is running on your network
  • Open source (MIT license, 235 stars on GitHub)

The extension works with AnythingLLM v1.6.5 and later. It supports both Desktop app and Docker deployments. If you're using an older version, update first.

Which Browsers Does the Extension Support?

The extension runs on all major desktop browsers: Chrome, Microsoft Edge, Brave, and Firefox. Installation process is identical across all of them.

Chrome, Edge, and Brave all use the same code base. Install from the Chrome Web Store and the extension works in all three. Firefox uses a separate Firefox Add-ons build but the functionality is identical. If you use multiple browsers, install it in each one--configuration is per-browser, so you'll need to connect each instance separately.

Android and mobile browsers are not supported. Related searches mention Android, but the extension is desktop-only. If you need mobile access to AnythingLLM, use the main app instead.

How to Install the AnythingLLM Browser Extension

Installation takes 90 seconds.

Step 1: Download from your browser's store

Chrome, Edge, or Brave: Visit https://chromewebstore.google.com/detail/anythingllm-browser-compa/pncmdlebcopjodenlllcomedphdmeogm and click "Add to [Browser]." A permissions dialog appears immediately.

Firefox: Search "AnythingLLM" in Firefox Add-ons and click "Add to Firefox."

Step 2: Grant permissions

The extension needs permission to read page content so it can capture text and HTML. This is required. Click "Allow" when the permission popup appears. You cannot use the extension without granting these permissions.

Step 3: Confirm the toolbar icon

After installation, the extension icon appears in your browser toolbar. It looks like a small chat bubble with "ALM" inside it. Click it to open the extension popup. You'll see a large "Connect" button and a text field labeled "Connection String."

Step 4: Test the context menu

Right-click on any webpage. If you see an "AnythingLLM" option in the context menu, installation is complete. You're ready to connect in the next step.

If the context menu option doesn't appear, reload the extension: go to your browser's extension settings, find AnythingLLM, and click the refresh icon. Close and reopen the browser if reloading doesn't work.

How to Connect the Extension to Your AnythingLLM Instance

This is the step that confuses most people. The extension needs two things: a connection URL and an API key. You generate the key in AnythingLLM, then paste both into the extension.

Step 1: Generate an API key in AnythingLLM Desktop

  1. Open the AnythingLLM Desktop app
  2. Click the gear icon (Settings) in the bottom left corner
  3. Navigate to the "Tools" tab
  4. Scroll down until you see "Browser Extension"
  5. Click the "Generate API Key" button
  6. A key appears. Copy it immediately--it's a long alphanumeric string. Treat it like a password and keep it private.

If you don't see the "Generate API Key" button:

  • Restart AnythingLLM
  • Go back to Settings and refresh the page
  • Ensure you're running v1.6.5 or later (check the about screen)

Step 1 (Docker alternative):

  1. Open your Docker-hosted AnythingLLM instance in a browser on port 3001
  2. Click the gear icon (Settings) in the bottom left
  3. Go to "Tools"
  4. Scroll to "Browser Extension"
  5. Click "Generate API Key" and copy the key

Step 2: Create your connection string

Your connection string combines the AnythingLLM URL and the API key in this format:

https://<your-anythingllm-url>:<port>?apiKey=<your-api-key>

For a local instance, this typically looks like: your local machine's address or domain with port 3001, plus the full API key you copied. If AnythingLLM is running on a remote server, substitute the server's domain or IP address. Make sure you include the full key--if the key is missing or truncated, the connection fails.

Step 3: Paste into the extension

Open the extension popup by clicking its icon in the toolbar. Paste your full connection string into the text field. Click "Test Connection." A green checkmark appears if it works. A red error appears if something is wrong.

If the test fails:

  • Verify AnythingLLM is actually running and accessible at that URL (try opening the URL in a new browser tab)
  • Double-check the API key--copy it again from Settings if you're not sure
  • Make sure the entire URL is included, with no extra spaces

Manual connection vs. auto-registration:

If AnythingLLM is running locally on your machine, the extension sometimes offers an "Auto-Register" button instead. Clicking it automatically detects your instance and generates a key. This is faster if it works. But if auto-registration fails or your AnythingLLM is on a remote server, manual connection (the method above) is always available and always works.

Using the Extension: Save Text and Pages

Once connected, right-click any webpage to see your options:

  • Save Selection to AnythingLLM: Highlight text on a page, right-click, and choose this option. The text is sent to your default workspace and indexed as searchable content.
  • Embed Full Page in AnythingLLM: Right-click anywhere on the page and choose this option to capture the entire page. Every element--text, headings, links, structured data--is indexed.
  • Save Link to AnythingLLM: Saves the URL of the current page as a reference.

When you choose an option, a small popup appears asking which workspace to send the content to. This is a unique feature. Generic web clippers just dump content into a single collection. AnythingLLM's extension lets you organize captures by workspace at the moment of saving. Your customer support team can send documentation to the "Support RAG" workspace. Your product team can send design specs to "Product Design." No extra steps, no manual sorting later.

Difference between snippet and full page:

Saving a snippet is fast and focused. You're capturing a specific piece of information. Embedding the full page is slower but more comprehensive--AnythingLLM indexes every element on that page, which means more context for retrieval but larger storage footprint. Use snippets for targeted information. Use full page embedding for documentation sites, wikis, and reference material you want fully searchable.

Can I Send Password-Protected or Hidden Content?

Yes. This is where privacy becomes a real advantage.

Password-protected pages: If you're logged into a SaaS tool--HubSpot, Pipedrive, your internal company intranet, anything behind a login--the extension can capture what you see. Because the extension runs locally in your browser, it only accesses content your browser has already rendered. It doesn't send credentials to AnythingLLM or any third party. The content goes directly from your screen to your local AnythingLLM instance.

VPN-locked pages: Same principle. If you can access a page through a VPN, the extension can capture it. The extension respects your network setup. It doesn't bypass VPN or attempt to reach content externally.

Data handling: The extension never sends content to Mintplex Labs, Opsily, or any cloud service. Everything is sent directly to your AnythingLLM instance, which you control. Review the privacy policy on the Chrome Web Store for full details, but the short answer: your proprietary content stays in your infrastructure. If your AnythingLLM is running on an air-gapped network, the extension can't reach it--but equally, no data leaves that network.

This is why teams use the extension for sensitive material. Customer data, internal process documentation, confidential research--all can be captured and indexed without risk of exposure to external services. The extension is just a bridge from your browser to your own instance.

Troubleshooting Common Connection Issues

Error: "Failed to connect" or timeout

The extension can't reach your AnythingLLM instance. Debug:

  1. Verify AnythingLLM is actually running and accessible at your connection URL (try opening it in a new browser tab)
  2. If you're using Docker, run docker ps in a terminal. Is the AnythingLLM container listed and showing "running"? If not, start it.
  3. If AnythingLLM is on a remote server, verify your network can reach that server and port 3001 is not blocked by a firewall.
  4. Check the connection string in the extension. Are there extra spaces? Is the API key complete? Copy the key fresh from Settings if you're unsure.

Auto-register button doesn't appear or fails

Auto-register only works if AnythingLLM is running locally and discoverable. If it fails:

  1. Use manual connection instead. Copy the full connection string (URL + API key) and paste it directly into the extension.
  2. If AnythingLLM is on a different machine, auto-register won't detect it. Manual connection is required.

Missing "Generate API Key" button in Settings

  1. Restart AnythingLLM completely
  2. Go back to Settings > Tools and refresh the page
  3. Verify you're running AnythingLLM v1.6.5 or later (check the about screen)
  4. If using Docker, pull the latest image and restart the container

Right-click context menu doesn't show AnythingLLM option

  1. Reload the extension: go to your browser's extension settings, find AnythingLLM, click the refresh icon
  2. Close and reopen the entire browser
  3. Verify permissions were granted during installation. Check your browser's extension settings and ensure AnythingLLM has permission to read page content.
  4. If still missing, uninstall and reinstall the extension

Content doesn't appear in workspace after saving

  1. Check that you selected the correct workspace in the extension popup before saving
  2. Verify the connection is still active: open the extension and confirm the green checkmark is visible
  3. Refresh your AnythingLLM workspace to see new content
  4. Check AnythingLLM's error logs if content still doesn't appear (indicates a backend issue)

Frequently Asked Questions

How do I install the AnythingLLM browser extension?

Go to the Chrome Web Store (or Firefox Add-ons), search "AnythingLLM," and click "Add to [Browser]." Grant permissions, and the extension installs in seconds. Right-click on any webpage to confirm it's working.

Where do I get the connection string and API key?

Open AnythingLLM, go to Settings > Tools > Browser Extension, and click "Generate API Key." Copy that key and combine it with your AnythingLLM URL (typically on port 3001) to create your connection string, then paste it into the extension.

Does the extension work offline?

No. AnythingLLM must be running and accessible. If you lose network access to your AnythingLLM instance, the extension can't send content.

What browsers are supported?

Chrome, Microsoft Edge, Brave, and Firefox (desktop). Mobile browsers are not supported.

How do I connect if auto-register fails?

Manual connection always works. Paste your full connection string (URL + API key) into the extension's text field and click "Test Connection."

Can I send password-protected pages?

Yes. The extension captures whatever your browser can see, including gated content, VPN-locked pages, and internal tools. Content is sent only to your AnythingLLM instance, never to third parties.

Does the extension work with Desktop and Docker?

Both. It works with Desktop app, Docker containers, and any deployment that exposes the AnythingLLM API. Connection setup is identical.

What's the difference between saving a snippet and embedding a full page?

A snippet captures the text you select. Full page embedding captures the entire HTML. Snippets are fast and targeted. Full page embedding is slower but indexes every element, giving more context for retrieval.

The Bottom Line

The AnythingLLM browser extension closes the gap between web browsing and knowledge management. Capture customer dashboards, competitor research, internal documentation, and API references with a right-click. Setup takes 10 minutes, and there's no friction after that.

You're running AnythingLLM to build a private RAG system. The extension makes that system live by feeding it current information directly from the web, without any third-party exposure. Install it across your team's browsers, connect it once, and you're capturing knowledge as you work.

If you're building this seriously, the next step is moving AnythingLLM out of your laptop and into managed hosting. Opsily hosts AnythingLLM so you focus on building your RAG pipeline, not managing infrastructure. See AnythingLLM pricing and deployment options to understand the full cost and how managed hosting compares to self-hosting.

Run AnythingLLM Without Setup Burden
Opsily handles infrastructure updates so you build RAG workflows instead of managing servers.
Get Started Free

Ready to self-host your own apps?

One server. Multiple apps. No per-app fees.

Get started →