Ask Grok on X: how to tag @grok and get a useful answer

How to ask @grok, step by step

Asking Grok in a reply takes a few seconds. X added the feature in March 2025.

1. Find the post you want to ask about, on x.com or in the X app.
2. Tap Reply. You can reply to the original post or to any reply further down the thread.
3. Type @grok followed by your question, for example: “@grok what is the source for this chart?”
4. Post the reply. Grok answers with its own public reply in the same thread.
5. Ask a follow-up by replying to Grok’s answer and tagging @grok again.

Keep questions short and specific. “Is this claim accurate, and what is the original source?” works better than “thoughts?”

What Grok reads before it answers

When you tag @grok, it reads the parent post and the thread above it. That is the context it uses to understand what you are asking about. If the important detail sits in a different thread or in a linked article, say so in your question.

Grok answers fast, but it can be wrong. Like any AI model, it can produce confident answers that are false, known as hallucinations. Grok’s news summaries on X once presented a false story as real, in April 2024. Check anything important against the original source.

Your question and the answer are public

An @grok reply is a normal public post. Anyone who can see the thread can see your question and Grok’s answer, and they may be shared or quoted.

• Do not include personal details, account numbers or anything you would not post publicly.
• For private questions, use the Grok tab instead. Chats there are private to you.
• Your public posts and Grok interactions can be used for training unless you opt out under Settings, then Privacy and Safety, then Data sharing and personalization, then Grok and Third-party Collaborators.

Our Grok privacy guide covers the opt-out and history deletion in detail.

Who can ask @grok right now

X launched @grok replies in March 2025. Around March 2026, @grok thread replies were reportedly limited to Premium and Premium+ subscribers. X made no formal announcement, and this comes from third-party reporting.

On October 12, 2026, SpaceXSI (formerly xAI) launched the X Subscription, with Free, Plus ($8), Premium ($30), Super ($100) and Ultra ($200) tiers. It replaced X Premium for new buyers. How the reported @grok limit maps to the new Free and Plus tiers is unconfirmed. The Free tier includes “limited chat requests” that reset weekly, but SpaceXSI publishes no numbers.

Good and bad uses

Where @grok helps

• Getting quick background on a news post or a trending topic.
• Explaining jargon, acronyms or a meme in a thread.
• Summarizing a long thread you do not have time to read.
• Asking for the likely original source of a claim, then checking it yourself.
• Translating a post written in another language.

Where it does not

• Settling a factual dispute on its own word. Grok can be wrong, and a public wrong answer spreads.
• Medical, legal or financial decisions. Use a qualified professional.
• Anything involving private information about you or other people.
• Piling on in arguments. Tagging Grok to dunk on someone rarely adds information.

Image requests through @grok

Image requests via @grok have the strictest rules, because of the deepfake scandal of December 2025 and January 2026. Users had @grok make nonconsensual sexualized edits of real people, including minors, and regulators in several countries opened investigations.

• January 9, 2026: image replies on X were limited to paid users.
• January 14, 2026: Grok on X was barred from editing images of real people into revealing clothing.
• Around March 2026: image generation and editing through @grok were reportedly limited to Premium and Premium+ subscribers.

For general image and video creation, the standalone tool is Grok Imagine. For everything else Grok does on the platform, see our Grok on X overview.

Questions people ask