Grok Prompts

xAI · text · Website ↗

Grok is xAI's model with built-in access to the live X feed and the web. That changes which tasks are worth giving it: it answers not only from training data but from what is happening right now.

This page holds two kinds of prompts. Some work only with Grok — trend and thread analysis on X, brand-reputation monitoring, finding a news hook — because they need the live feed that other models cannot reach. The rest are the library's general prompts, which Grok handles on par with ChatGPT, Claude and Gemini.

For anything tied to current data, state the period explicitly ("over the past three days") and ask the model to show what the conclusion rests on — specific posts or sources. Without that you get a confident summary instead of a verifiable answer.

Personal Development & Hobbies

SMART Goal Setting Coach

Coaches you through turning any ambition into a SMART goal with milestones, obstacle plans, and accountability.

{} 8 vars
Marketing

SEO Content Optimizer & Keyword Researcher

Researches keywords and optimizes content end-to-end for a target keyword: titles, meta, structure, and SEO fixes.

{} 5 vars
Programming & Development

Professional Code Generator

Generates clean, well-documented, production-ready code from your requirements, following best practices in any language.

{} 6 vars
Creative

Master Storyteller

Crafts an original, emotionally resonant story tailored to your chosen genre, audience, theme, and length.

{} 7 vars
Social Media & Communication

LinkedIn Professional Post Creator

Writes a polished, engagement-optimized LinkedIn thought-leadership post that builds your professional credibility.

{} 8 vars
Business

Comprehensive Business Plan Builder

Guides an AI to build a complete, investor-ready business plan section by section from your idea.

{} 6 vars
Travel

Hotel Vetting Before You Book

Vets a hotel before you pay: separates checkable facts from marketing copy, finds recurring red flags in the reviews, reads the cancellation terms and returns a verdict with caveats.

{} 5 vars
Entertainment & Games

D&D 5e Character Creation Wizard

A guided D&D 5e character creator that walks you through race, class, ability scores, background and spells one step at a time — and ends with a complete, table-ready sheet.

{} 5 vars
Social Media & Communication

Telegram Bot Conversation Designer

Designs complete Telegram bot conversations: onboarding, scenario flows built on inline keyboards, real message copy, error recovery paths, and a command menu.

{} 5 vars
Social Media & Communication

Telegram Channel Content Plan Builder

Builds a 30-day Telegram channel content plan with pillars, post formats, opening hooks, interactive posts, and CTAs designed around forwards and notifications.

{} 5 vars
Social Media & Communication

YouTube Video Title & Description Optimizer

Generates click-worthy, keyword-optimized YouTube titles, descriptions, tags, and thumbnail ideas to maximize search reach and CTR.

{} 17 vars
Food & Cooking

Weekly Meal Planner

Builds a structured 7-day meal plan with recipes, shopping list, and prep tips.

{} 9 vars
IT & Administration

Ansible Automation Playbook Creator

Generates production-ready, idempotent Ansible playbooks and roles for any infrastructure automation or configuration task.

{} 5 vars
Marketing

Brand Voice & Messaging Strategy Developer

Builds a complete brand voice, tone, and messaging strategy with frameworks, hierarchy, and ready-to-use examples.

{} 8 vars
Creative

Professional Content Editor

Expert multi-level editing and proofreading with structured, constructive feedback on any text you submit.

{} 6 vars
Travel

Adventure Travel Planner

Builds a safe, day-by-day adventure travel itinerary with gear, fitness, logistics, and risk guidance.

{} 6 vars
Programming & Development

Advanced Debugging Assistant

Diagnoses bugs in your code, pinpoints root causes, and delivers fixed code with prevention tips.

{} 6 vars
Creative

Poetry Workshop Facilitator

An expert poetry mentor who critiques your draft on craft and feeling, then teaches you to revise it.

{} 6 vars
Entertainment & Games

Movie Recommendation Engine

Generate a personalized, spoiler-free movie and TV watchlist tailored to your tastes, mood, and streaming platforms.

{} 8 vars
Social Media & Communication

Community Management Response Templates

Builds a brand-voiced library of social media reply templates for every common interaction and scenario.

{} 11 vars
Programming & Development

Code Review Expert

Performs a rigorous senior-level code review covering correctness, design, performance, and security with prioritized, actionable fixes.

{} 5 vars
Marketing

PPC Campaign Generator & Ad Copy Creator

Generates high-converting, platform-specific PPC ad copy and campaign structure for any product, audience, and budget.

{} 10 vars
Marketing

Competitor Analysis & Market Research Tool

Runs a structured competitor and market analysis that surfaces gaps and strategic recommendations for your business.

{} 8 vars
IT & Administration

Cloud Infrastructure Architect

Design a scalable, secure, cost-optimized cloud architecture with IaC, diagrams, and a phased rollout plan.

{} 8 vars

Writing prompts for Grok

How Grok differs from ChatGPT, Claude and Gemini

The one difference that actually changes how you frame a task is access to the live X feed. On everything else Grok sits in the same range as its three neighbours and handles the library's general prompts on par with them.

Which gives a simple rule for choosing: if the task contains no "right now", "today" or "in this thread", the model matters little and Grok offers no edge. If it does, the other three simply cannot do the job.

Rewriting a library prompt for Grok

  • State the window explicitly. "Over the last 72 hours", "since Monday". Without a period the model quietly picks one, and running the same prompt again gives a different picture.
  • Demand the evidence. "Cite 5–7 specific posts the conclusion rests on." It is the only way to tell an observation from a retelling.
  • Separate claim from fact. A feed is what people are writing, not what happened. Phrasing the ask as "what is being claimed" rather than "what happened" saves an hour of checking later.

Where it falls down

It reads loud as widespread: a dozen visible posts easily become a "trend". Ask it to estimate scale — roughly how many accounts, how it spreads over time — and treat the number as an order of magnitude.

It picks up the register of the feed. If the underlying discussion is combative, the answer comes out combative even where the prompt asked for a neutral read; set the tone on its own line.

And it separates activity around a topic from the topic's importance poorly: a spike can turn out to be two accounts arguing. If the conclusion feeds into real work, verify it against sources outside X.