Model
Grok in any tab: Grok 4.7 in a browser sidebar
Copilotly runs xAI's Grok 4.7 in a sidebar beside the page you are on, in Chrome, Edge, Firefox, Brave and Opera. It is not made by xAI or X; it offers Grok alongside GPT, Claude and Gemini.
By the Copilotly editorial team. Last verified November 11, 2026 against the current Copilotly release.
Quick answer
To use Grok on any website without opening X, install the Copilotly extension, choose Grok 4.7 in the model menu and press Alt+C. The free plan includes Grok 4.20 Fast and 10 Grok 4.7 messages a day. xAI's API docs also list a 500K-token context for Grok 4.7. Copilotly is independent of xAI.
- Grok 4.7 and the fast Grok 4.20 in the model menu
- No need to open X or a separate site
- Free plan includes 10 premium-model messages a day
- Or use your own xAI API key, unmetered by Copilotly
At a glance
- Grok models
- Grok 4.7 (premium), Grok 4.20 Fast (fast)
- Grok 4.7 context
- 500K tokens, text and image input (per xAI, Nov 2026)
- Price to start
- Free, no card; 10 premium messages a day
- Works in
- Chrome, Edge, Firefox, Brave, Opera, Arc, Vivaldi
- Your own key
- xAI API key, never metered by Copilotly
- Made by
- Copilotly, not xAI or X
How to use Grok in any tab
Install the extension
Add Copilotly from the download page for Chrome, Edge, Firefox, Brave, Opera, Arc or Vivaldi. Create a free account when it asks; there is no card.
Pick Grok 4.7 in the model menu
Open the model menu and choose Grok 4.7, described as witty, current and direct. Grok 4.20 Fast is the quick model, with a larger context window.
Open it on any page with Alt+C
Press Alt+C to open the sidebar beside whatever page you are on. Ask about the page, select text and rewrite it, or start a new chat. Switching to another model mid-conversation is one click in the same menu.
What does Grok in a browser sidebar do?
Choose Grok 4.7 and the sidebar answers with Grok while it can read the page you are on.
A direct take on what you are reading
Use it for a direct take on an article, a quick explanation, or to compare against another model on the same question. xAI's own guidance on its models page is "For everything else, including code, use Grok 4.7," and it calls Grok 4.7 the most capable model it has built.
Text and images
xAI's docs list text and image input for Grok 4.7, so you can ask about a chart or screenshot on the page. It also lists reasoning with adjustable effort and structured outputs on the API.
The rest of the product
The rest is the same whichever model answers: rewrite selected text in place, web search with cited sources, deep research and image generation.
Is there an official Grok browser extension?
We have not found one from xAI. as of November 2026, a web search turns up several third-party extensions with Grok in the name, and at least some state in their listings that they are not affiliated with xAI or X. xAI itself offers Grok on its own site and apps and inside X; see xAI's Grok page for what is available today and on which plans.
When xAI's own apps are the better choice
If you want Grok's tight connection to X posts and xAI's own features, use those.
Where Copilotly differs
- Grok on any website. It sits in a sidebar next to the page, not inside X.
- Every model in one menu. GPT, Claude, Gemini, DeepSeek and Mistral too.
- Any browser that runs the extension, including Firefox.
- Bring your own key. Use an xAI API key and pay xAI directly.
What Copilotly cannot do with X
Copilotly does not read your X feed and has no special access to X or to live X posts. For current information it uses web search with cited sources.
Which Grok models do you get, and how big is the context?
xAI's models page lists Grok 4.7 with a 500K-token context and the Grok 4.20 non-reasoning model with 1M (as of November 2026).
| Model | Tier | Free plan | Pro plan |
|---|---|---|---|
| Grok 4.20 Fast | fast | Included, within your token allowance | Included, within your token allowance |
| Grok 4.7 | premium | 10 messages a day, then Grok 4.20 Fast | Unlimited |
There is no frontier-tier Grok model in the list. Pro is $6.99/week, $19.99/month or $149.99/year, with a 7-day free trial for first-time subscribers.
Grok 4.7
Premium. xAI lists 500K tokens of context, text and image input, and API prices of $2.00 per million input tokens and $6.00 per million output tokens for prompts under 200K tokens, doubling above that. Free gets 10 messages a day.
Grok 4.20 Fast
Copilotly's quick model. xAI's page lists the 4.20 non-reasoning model with a 1M-token context, text and image input, at $1.25 per million input tokens and $2.50 per million output tokens under 200K tokens. It is available on every plan within your token allowance and is where Free falls back after the 10 daily premium messages.
The API prices are what you would pay xAI on your own key. Copilotly's plans are separate; see pricing.
Can I use Grok without an X subscription?
On the free plan
Yes, in Copilotly. The free plan includes Grok 4.20 Fast and 10 premium-model messages a day for Grok 4.7. You do not need a paid plan from X, and a free Copilotly account does not need a card.
On your own xAI API key
Every model on this page can also run on your own API key. Requests made on your own key are never metered by Copilotly, whatever the model's tier, and that includes frontier models on the free plan, which are otherwise Pro-only. You pay the provider directly at their API rates. The steps are on the bring your own key page.
With an xAI API key, Grok 4.7 runs unmetered in the sidebar and xAI bills you directly. xAI's API billing is separate from X subscriptions, so an X plan does not give you an API key.
When should I pick Grok 4.7 or Grok 4.20 Fast?
A different voice from the other models
Pick Grok 4.7 when you want a direct, conversational answer or a second opinion from a different provider. It is a reasoning model with a 500K-token context, which is plenty for most pages and articles.
Very long inputs and speed
For very long inputs or fast replies, Grok 4.20 Fast has the larger window by xAI's listing and is available on every plan.
Where accuracy matters
Check the answer against its sources, using web search with citations to see where a claim comes from. For a long task, deep research builds a sourced report instead of one answer.
How does Grok compare with the other models in the picker?
Grok 4.7 has the smallest context window of the flagships below, at 500K tokens by xAI's listing, which is still large for a web page or a long article.
| Model | Context window | Image input | Source (Nov 2026) |
|---|---|---|---|
| GPT-6.1 Terra | 1.05M | Yes | OpenAI docs |
| Claude Sonnet 5.5 | 1M | Yes | Copilotly catalogue |
| Gemini 4 Argon | 1M | Yes, plus audio, video and PDF | Google AI Studio docs |
| DeepSeek V4 Pro | 1M | No (Flash: yes) | DeepSeek docs |
| Grok 4.7 | 500K | Yes | xAI docs |
| Mistral Large 4 | 1M | Multimodal | Mistral docs |
When context size matters
It matters when you load a very large file or a long thread. For ordinary pages and articles, 500K tokens is far more than you need, so the model's quality on your task is the better guide. For the largest inputs, Grok 4.20 Fast lists 1M.
What is distinctive about Grok
Copilotly's picker describes Grok 4.7 as witty, current and direct. Those are descriptions of tone, not measurements. xAI's documentation positions it as its most capable model and the default for code. The difference you will notice, if any, is style, so judge it on a few of your own questions.
How do I get better answers from Grok in the sidebar?
Ask for sources on anything current
Grok's connection to X is xAI's, not Copilotly's. In Copilotly, turn on web search and ask for sources so you can open what the answer is based on.
Ask for the opposing view
A direct style can read as more certain than it is. Ask "What is the strongest case against this?" and "What would change your answer?"
Use images for what is on the page
xAI lists image input for Grok 4.7, so you can ask about a chart or screenshot. Check any figure it reads off a chart against the chart.
Compare across providers
On anything contested, ask Claude or GPT the same question. Differences show you where to dig.
Can Grok in Copilotly see X posts or live data?
No. This is the question people most often get wrong about third-party Grok tools, so it is worth being exact.
What stays with xAI
Grok on X has access that xAI controls, such as its connection to X posts. Copilotly reaches Grok through xAI's API, like any developer, and has no special access to X, your feed or live posts.
What you can do instead
- Turn on web search for current information, with numbered sources you can open.
- Use deep research for a broader sourced report on a topic.
- Open the X post in a tab and ask Grok about it with the sidebar, since the sidebar reads the page you are on.
Limits to remember
A page you can read in your own browser is what the sidebar sees. Content behind a login you do not have, or that has not loaded, is not available to it.
Does Grok work in Firefox, Opera and Safari?
Grok in Firefox comes from the Firefox Add-ons listing, and Grok in Opera from the Chrome Web Store build, which Opera, Brave, Arc and Vivaldi all install.
| Where | Status | How |
|---|---|---|
| Chrome | Supported | Chrome Web Store |
| Edge | Supported | Its own Edge Add-ons listing |
| Firefox | Supported | Its own Firefox Add-ons listing |
| Brave | Supported | Installs the Chrome Web Store build |
| Opera | Supported | Installs the Chrome Web Store build |
| Arc, Vivaldi | Supported | Install the Chrome Web Store build |
| Safari | Not yet | There is no Safari build in the App Store |
| iPhone, Android | Supported | The Copilotly iOS and Android apps, which work on the free plan |
Who it is for
Readers who want a quick, direct summary
On a long news or opinion article, open the sidebar and ask Grok 4.7 for the argument in three sentences and the strongest counterpoint. Then turn on web search and ask where the key claim comes from. A fluent summary is not a fact-check.
Analysts comparing models
Ask Grok 4.7 and one other model the same question about a page or file. Differences in what each flags are a prompt to read the source yourself. The model menu lets you swap without losing the chat.
People with an xAI API key
Connect the key under bring your own key to use Grok 4.7 beside any page without Copilotly metering, paying xAI's per-token rates directly.
Grok through xAI's apps or through Copilotly
| xAI's Grok apps and X | Grok in Copilotly | |
|---|---|---|
| Models | xAI's models | Grok plus GPT, Claude, Gemini, DeepSeek, Mistral |
| Connection to X | Built in by xAI | None |
| Browser extension | We found no official one | Sidebar in Chrome, Edge, Firefox, Brave, Opera |
| Works on any website | Copy and paste | Yes, Alt+C reads the open tab |
| Your own API key | Not how it is used | Yes, unmetered by Copilotly |
| Free to try | Check xAI's plans | Free plan, 10 premium messages a day |
If something goes wrong
Grok does not know about a recent post or event
Copilotly has no access to X or to live posts. Turn on web search for current information, and open the sources it cites.
The chat moved to Grok 4.20 Fast
That is the fallback after your 10 daily premium messages on Free. The count resets daily; requests on your own xAI key are never metered.
I want the biggest possible context window
Pick Grok 4.20 Fast, which xAI lists at 1M tokens against 500K for Grok 4.7. Models differ in how well they use very long inputs, so check the result against the source.
Alt+C does not open the sidebar
Another extension or the browser may be using the same shortcut. In Chrome, Edge and Brave open chrome://extensions/shortcuts (Edge uses edge://extensions/shortcuts) and check what is assigned to Copilotly. In Firefox, open the Add-ons page, use the gear menu and choose Manage Extension Shortcuts.
Guides that go with this
ChatGPT vs Claude vs Gemini: We Tested Real Legal, Medical, and Tax Questions
We put ChatGPT, Claude, and Gemini through dozens of real professional questions covering contract review, lab results, retirement planning, and tax deductions. Each platform showed clear strengths and surprising failure modes, and none of the 3 matched domain-specific copilots for high-stakes guidance.
Best AI Copilots by Profession (2026): Honest Picks for 8 Fields
We picked the best AI copilot for 8 professions and use cases in 2026: GitHub Copilot for developers, Microsoft 365 Copilot for office workers, and domain-specific copilots for legal, health, tax, finance, career, and small business questions. Every pick includes real alternatives, honest pricing, and the cases where a different tool wins.
Common questions
Is this the official Grok extension?
No. Copilotly is independent and not affiliated with xAI or X. It uses Grok models through xAI's API. We have not found an official xAI browser extension.
Which Grok models are available?
Grok 4.7 (premium) and Grok 4.20 Fast (fast).
Can I use Grok without an X subscription?
Yes, in Copilotly. The free plan covers Grok 4.20 Fast and 10 premium-model messages a day, or you can use your own xAI API key. A free Copilotly account is required; a card is not.
Does it have access to live X posts?
No. Copilotly has no special access to X. For current information it uses web search with cited sources.
What is Grok 4.7 best at?
xAI's docs describe Grok 4.7 as the most capable model it has built and recommend it for code and everything else. Copilotly's picker calls it witty, current and direct. For your own work, try the same question on it and one other model.
Can Grok read images?
xAI's docs list text and image input for Grok 4.7, so you can ask about a chart or screenshot. Verify any figure it reads against the image.
Can I use Grok in Firefox, Opera or Safari?
Copilotly is the same extension across browsers, so Grok is too. It is listed for Firefox and Edge. Brave and Opera install the Chrome Web Store build, and Arc and Vivaldi do the same. Safari is not supported yet: there is no Safari build in the App Store, so use one of the other browsers on a Mac, or the iOS and Android apps on a phone.
Sources
Keep going
Related
More ai models
Use Grok on any page
Install the free extension, pick Grok 4.7 and press Alt+C. No card needed.