These are prompt skeletons, not one-off examples. Each one is tested structure with [BRACKET] placeholders you replace. Copy a template, swap the brackets for your specifics, delete the ones you do not need, and send. The wording around the brackets is the part that makes Grok behave — keep it.

Every text template here follows the same backbone: [Role] + [Task] + [Context] + [Constraints] + [Output format]. Image and video templates follow subject + environment + lighting + style + camera/technical. Once you internalize those two shapes you can build your own. For a shorter reference, pair this page with the Grok prompt cheat sheet, and see the full best Grok prompts roundup for finished, non-template versions.

Advertisement

How to use these templates

Replace each [BRACKET] with a real value before sending — do not leave literal brackets, or Grok will stop and ask what you meant. Three rules keep the results sharp:

  • Fill or cut. If a bracket does not apply, delete the whole phrase, not just the bracket.
  • Name the mode. Add "Use DeepSearch," "Think step by step," or "Check X" so Grok picks the right tool.
  • Lock the output. The [OUTPUT FORMAT] bracket is the highest-leverage one. Be specific: "a markdown table," "300 words," "valid JSON."

Real-time X monitoring

These trigger Grok's live X search. Lead with "Check X" or "What are people saying on X about" so Grok reads current posts instead of memory. For a deeper walkthrough, see the roundup.

1. Topic pulse check

Check X for what people are saying about [TOPIC] in the last [TIME WINDOW, e.g. 24 hours]. Summarize the [NUMBER] most common viewpoints, note the overall sentiment (positive/negative/mixed), and flag any claim that is disputed. Ignore spam and giveaways. Output: a bulleted brief with one representative quote per viewpoint and a link to each source post.

Swap: [TOPIC], [TIME WINDOW], [NUMBER].

2. Brand & mention monitor

Check X for posts mentioning [BRAND OR HANDLE] from the last [TIME WINDOW]. Separate them into praise, complaints, questions, and news. For complaints and questions, note whether they need a reply and suggest a one-line response. Exclude retweets with no added comment. Output: a table with columns Category, Post summary, Author, Suggested action.

Swap: [BRAND OR HANDLE], [TIME WINDOW].

3. Breaking-event tracker

What are people saying on X about [EVENT] right now? Build a timeline of what is confirmed versus rumored over the last [TIME WINDOW]. Prioritize posts from [TYPE OF ACCOUNT, e.g. journalists / officials / eyewitnesses] and clearly mark anything unverified. Output: a chronological timeline, then a short "what we still do not know" section.

Swap: [EVENT], [TIME WINDOW], [TYPE OF ACCOUNT].

4. Competitor & trend watch

Check X for what [COMPETITOR OR NICHE] accounts have posted in the last [TIME WINDOW] about [SUBJECT]. Identify emerging trends, product hints, and any post gaining unusual traction. Tell me what my [MY ROLE/COMPANY] should pay attention to. Output: 5 bullets of signal, each with the source post and why it matters.

Swap: [COMPETITOR OR NICHE], [TIME WINDOW], [SUBJECT], [MY ROLE/COMPANY].

DeepSearch research

These force DeepSearch, which pulls and cites live sources rather than guessing. Always keep the "cite every source" line. For research-first prompts, also see the full roundup.

5. Grounded question with sources

You are a [DOMAIN] research analyst. Use DeepSearch to answer: [QUESTION]. Pull from at least [NUMBER] independent, recent sources and cite every one with a link. Distinguish established facts from contested claims, and note the date of each key figure. Output: a direct answer first, then supporting evidence, then a "confidence and gaps" note.

Swap: [DOMAIN], [QUESTION], [NUMBER].

6. Compare options / buyer research

Use DeepSearch to compare [OPTION A] vs [OPTION B] vs [OPTION C] for [USE CASE / BUYER]. Evaluate on [CRITERIA, e.g. price, performance, support, lock-in]. Cite current pricing and specs with source links and dates. Recommend one and explain the trade-off. Output: a comparison table, then a one-paragraph recommendation for [BUDGET / CONSTRAINT].

Swap: [OPTION A/B/C], [USE CASE / BUYER], [CRITERIA], [BUDGET / CONSTRAINT].

7. Literature / evidence review

Use DeepSearch and Think mode. Review the current evidence on [CLAIM OR TOPIC] published since [DATE]. Summarize what the strongest sources agree on, where they disagree, and what remains unsettled. Cite each source with a link and publication date, and weight studies over opinion pieces. Output: an executive summary, a "points of agreement" list, and a "points of contention" list.

Swap: [CLAIM OR TOPIC], [DATE].

Prefer a dedicated deep-dive? The roundup collects finished research prompts you can adapt.

Advertisement

Coding

Give Grok the language, the goal, and the constraints, then demand runnable output. For finished, copy-paste dev prompts, see the coding prompts pack.

8. Build a function or component

You are a senior [LANGUAGE] engineer. Write a [FUNCTION / COMPONENT] that [WHAT IT DOES]. Inputs: [INPUTS]. Output: [EXPECTED OUTPUT]. Constraints: [e.g. no external deps, must handle empty input, O(n) time]. Include inline comments and [NUMBER] unit tests covering edge cases. Output: one code block, then a two-line usage example.

Swap: [LANGUAGE], [FUNCTION / COMPONENT], [WHAT IT DOES], [INPUTS], [EXPECTED OUTPUT], constraints, [NUMBER].

9. Debug & explain

Think step by step. Here is [LANGUAGE] code that should [INTENDED BEHAVIOR] but instead [ACTUAL BEHAVIOR / ERROR]:

[PASTE CODE]

Find the root cause, explain why it happens in plain terms, and give the corrected code. List any other bugs you spot. Output: root cause, fixed code block, then a bullet list of remaining risks.

Swap: [LANGUAGE], [INTENDED BEHAVIOR], [ACTUAL BEHAVIOR / ERROR], [PASTE CODE].

10. Refactor to a standard

Refactor this [LANGUAGE] code to be [GOAL, e.g. more readable / faster / idiomatic] while keeping behavior identical. Follow [STYLE OR PATTERN, e.g. functional style, PEP 8, our naming convention]. Do not change the public interface. Explain each change in one line.

[PASTE CODE]

Output: the refactored code block, then a short changelog.

Swap: [LANGUAGE], [GOAL], [STYLE OR PATTERN], [PASTE CODE].

Writing & content

Set the role, audience, and format so Grok writes in the right register instead of generic prose.

11. Article or blog post

You are a [TYPE] writer for [AUDIENCE]. Write a [LENGTH]-word [FORMAT, e.g. how-to article] on [TOPIC]. Angle: [SPECIFIC TAKE]. Tone: [TONE]. Include [SPECIFIC ELEMENTS, e.g. an intro that leads with the answer, 3 subheads, one example]. Avoid [BANNED WORDS/CLICHES]. Output: the finished piece with subheadings, no preamble.

Swap: [TYPE], [AUDIENCE], [LENGTH], [FORMAT], [TOPIC], [SPECIFIC TAKE], [TONE], elements, banned words.

12. Social post / thread

Write [NUMBER] [PLATFORM] posts about [TOPIC] for [AUDIENCE]. Goal: [GOAL, e.g. drive signups / spark replies]. Voice: [VOICE]. Each post must stand alone, lead with a hook, and stay under [CHARACTER LIMIT]. No hashtag spam. Output: numbered posts, each ready to paste, with the strongest one marked.

Swap: [NUMBER], [PLATFORM], [TOPIC], [AUDIENCE], [GOAL], [VOICE], [CHARACTER LIMIT].

13. Rewrite & tighten

Rewrite the text below to be [GOAL, e.g. clearer / shorter / more persuasive] for [AUDIENCE]. Keep every fact, cut filler, and match a [TONE] tone. Target [LENGTH]. Do not add new claims.

[PASTE TEXT]

Output: the rewrite, then one line on what you changed and why.

Swap: [GOAL], [AUDIENCE], [TONE], [LENGTH], [PASTE TEXT].

Business & analysis

These turn Grok into an analyst. Give it the decision, the constraints, and the output document you actually need.

14. Decision brief

You are a [ROLE, e.g. strategy advisor] for a [COMPANY TYPE]. We are deciding whether to [DECISION]. Context: [KEY FACTS AND CONSTRAINTS]. Lay out the options, the main trade-offs, the biggest risk, and what you would do and why. Assume a [TIME/BUDGET] constraint. Output: a one-page brief — recommendation, rationale, risks, next step.

Swap: [ROLE], [COMPANY TYPE], [DECISION], [KEY FACTS], [TIME/BUDGET].

15. Data & document summary

Analyze the [DATA / DOCUMENT] below for [AUDIENCE]. Answer: [SPECIFIC QUESTIONS]. Pull out the [NUMBER] most important findings, quantify where possible, and flag anything surprising or missing. No speculation beyond the data.

[PASTE DATA / DOCUMENT]

Output: headline finding, key numbers, then a short "so what" for [AUDIENCE].

Swap: [DATA / DOCUMENT], [AUDIENCE], [SPECIFIC QUESTIONS], [NUMBER].

16. Meeting / email drafter

Draft a [EMAIL / MESSAGE] from me ([MY ROLE]) to [RECIPIENT] about [PURPOSE]. Key points to land: [POINTS]. Desired outcome: [OUTCOME]. Tone: [TONE]. Keep it under [LENGTH] and end with one clear ask. Output: subject line plus body, ready to send.

Swap: [EMAIL / MESSAGE], [MY ROLE], [RECIPIENT], [PURPOSE], [POINTS], [OUTCOME], [TONE], [LENGTH].

Grok Imagine image & video

Grok Imagine (Aurora for stills, plus ~10-second 720p video clips with audio) rewards the formula subject + environment + lighting + style + camera/technical. Keep prompts to 30–80 words, lead with the subject, and pick one coherent style — skip quality-adjective spam.

17. Image (Aurora)

[SUBJECT with 2-3 defining details], in [ENVIRONMENT / SETTING], lit by [LIGHTING, e.g. soft golden hour light], in the style of [ONE STYLE, e.g. editorial photography / cinematic 3D render], shot on [CAMERA / LENS or ANGLE, e.g. 50mm, low angle], [MOOD]. [ASPECT RATIO].

Swap: [SUBJECT], [ENVIRONMENT], [LIGHTING], [ONE STYLE], [CAMERA / LENS], [MOOD], [ASPECT RATIO].

18. Video clip

[SUBJECT] [DOING ONE CLEAR ACTION], in [ENVIRONMENT], [LIGHTING]. Camera [MOVEMENT, e.g. slow push-in / static]. Style: [ONE STYLE]. Mood and pacing: [MOOD]. Ambient audio: [SOUND]. Keep the motion simple and continuous for a ~10-second clip.

Swap: [SUBJECT], [ACTION], [ENVIRONMENT], [LIGHTING], [MOVEMENT], [ONE STYLE], [MOOD], [SOUND].

For more finished visual prompts and the reasoning behind each one, browse the best Grok prompts roundup.

Frequently Asked Questions

What is a Grok prompt template?

A prompt template is a reusable skeleton with [BRACKET] placeholders. You keep the tested structure — role, task, context, constraints, and output format — and swap only the brackets for your specifics. It turns a one-off prompt into something you can run every day.

How do I use the [brackets] in these templates?

Copy the template, then find-and-replace each [BRACKET] with your real value. Delete brackets you do not need and keep the surrounding wording. Do not leave literal brackets in the final prompt, or Grok will treat them as placeholders to ask about.

Which templates use DeepSearch?

The research templates lean on DeepSearch, which pulls and cites live sources instead of answering from memory. Add "Use DeepSearch and cite every source" to any research template to force grounded, linked answers.

How do I trigger real-time X answers?

Start the prompt with "Check X for" or "What are people saying on X about." That signals Grok to search live X posts. The real-time templates here are built around those triggers with brackets for the topic, time window, and account filters.

Do these work in the free version of Grok?

The text, coding, writing, and business templates work on any Grok tier. DeepSearch, Think, and Grok Imagine access depend on your plan; the heaviest parallel-agent work needs SuperGrok Heavy. The template structure is identical regardless of tier.

What is the standard formula behind these templates?

Text templates follow [Role] + [Task] + [Context] + [Constraints] + [Output format]. Image and video templates follow subject + environment + lighting + style + camera or technical detail, kept to 30 to 80 words with one coherent style.

Can I chain these templates together?

Yes. A common chain is a DeepSearch research template to gather grounded facts, then a writing template that references those findings, then a business template to turn the draft into a decision doc. Keep each step in the same conversation so Grok holds the context.

Advertisement