Skip to content
BotServBotServ
Claude PromptsClaude CommandsPrompt EngineeringAI PromptsClaude.aiAnthropicClaude Codes

100 Claude Prompts & Commands: The Ultimate List

100 Claude prompts across 10 categories: coding, marketing, SEO, video, copywriting, business, analysis and more with examples.

S

schutzgeist

6 min read
100 Claude Prompts & Commands: The Ultimate List

100 Claude Prompts & Commands: The Ultimate List for Claude.ai

What This Article Covers

  • 100 tested Claude prompts and command codes across 10 categories.
  • Slash-style shorthand commands (/optimize, /template, /matrix, /table) as prompt prefixes.
  • Tailored to Claude’s strengths: long context windows, XML tags, Artifacts, Claude Code.
  • Same categories as our ChatGPT guide, with Claude-specific tips.

Introduction

Claude.ai (Anthropic) responds to the same directive logic as ChatGPT: slash-style codes work identically. But Claude has particular strengths worth leveraging:

  • XML Tags: Claude is trained to parse <document>, <code>, <context> and similar tags exceptionally well, perfect for structured prompts.
  • Long Context: Feed it entire books or repositories, and /analyze commands become even more powerful.
  • Artifacts: Code, tables, and documents render as separate artifacts: /table and /document land there cleanly.
  • Claude Code: For coding tasks, the CLI variant (claude in your terminal) is the more capable option, and the same commands work there.

The list follows the same logic as ChatGPT, with 100 codes below plus Claude-specific refinements.

How These Commands Work in Claude

/optimize
<code>
[Your code here]
</code>

The XML tag trick is Claude-specific: wrap your input in <code>, <text>, or <data> tags and Claude processes it more cleanly than unmarked text.

Category 1: Coding & Development (10 Commands)

CommandWhat It DoesExample Prompt
/optimizeOptimize code”/optimize <code>[Code]</code>: focus on readability”
/debugFind and explain errors”/debug <code>[Code]</code> error: [message]“
/refactorRestructure code”/refactor to clean code, keep API intact”
/testWrite tests”/test with pytest, consider edge cases”
/documentAdd docstrings/docs”/document in Sphinx style”
/reviewCode review”/review for security + performance”
/explainExplain line by line”/explain for a junior dev”
/regexBuild regex pattern”/regex for date formats”
/sqlGenerate SQL”/sql for revenue per month in 2025”
/apiDesign endpoint”/api REST for user auth in FastAPI”

Claude tip: For coding work, use Claude Code (CLI), where Claude can directly read and write files and run tests. The same commands apply.

Category 2: Marketing & Advertising (10 Commands)

CommandWhat It DoesExample Prompt
/copyAd copy”/copy for [product], tone: warm”
/campaignCampaign concept”/campaign launch [product], 3 channels”
/taglineSlogans”/tagline 10 variations, concise”
/audienceBuild persona”/audience for [product] in detail”
/funnelFunnel stages”/funnel awareness→sale for [product]“
/ab-testVariations”/ab-test subject line [text], 5 variants”
/brand-voiceTone/voice”/brand-voice 5 adjectives + sample text”
/adAd copy”/ad for Meta Ads, hook + CTA”
/newsletterNewsletter”/newsletter weekly, 3 topics”
/launchLaunch plan”/launch 4-week plan for [product]”

Claude tip: Claude excels at longer, nuanced writing. For marketing copy with tone, it often digs deeper than ChatGPT.

Category 3: Video Editing & Content (10 Commands)

CommandWhat It DoesExample Prompt
/scriptVideo script”/script YouTube video [topic], 10 min, hooks”
/storyboardPlan scenes”/storyboard for 60s clip, shot list”
/hookOpening hook”/hook 5 variations for [topic]“
/thumbnailThumbnail idea”/thumbnail concept for [video]“
/chaptersChapters”/chapters from <transcript>[text]</transcript>”
/brollB-roll ideas”/broll list for [topic] video”
/shortsShort version”/shorts from this script”
/voiceoverVoiceover text”/voiceover 30s, calm, timing markers”
/title-videoTitle variations”/title-video 10 CTR-optimized titles”
/descriptionVideo description”/description SEO-optimized + tags”

Claude tip: With long transcripts in <transcript> tags, Claude can process entire videos. Perfect for /chapters and /summary.

Category 4: Writing & Copy (10 Commands)

CommandWhat It DoesExample Prompt
/rewriteRewrite”/rewrite [text], more concise”
/expandExpand”/expand with more detail”
/shortenShorten”/shorten to 50% length”
/toneChange tone”/tone → formal instead of promotional”
/proofreadProofread”/proofread spelling + style”
/simplifySimplify”/simplify for non-experts”
/headlineHeadlines”/headline 10 variations”
/introIntroduction”/intro for article on [topic]“
/conclusionConclusion”/conclusion with call-to-action”
/translateTranslate”/translate to German, natural”

Category 5: SEO (10 Commands)

CommandWhat It DoesExample Prompt
/keywordsKeyword ideas”/keywords for [topic] with intent”
/metaMeta tags”/meta title + description for [page]“
/outlineOutline”/outline H2/H3 for [keyword]“
/schemaSchema.org markup”/schema FAQ JSON-LD for [topic]“
/interne-linksInternal link suggestions”/interne-links for [article]“
/titleTitle variations”/title 5 CTR-optimized titles”
/snippetSnippet text”/snippet for [question], 40 words”
/content-gapFind gaps”/content-gap in <text>[article]</text>”
/eeatE-E-A-T improvements”/eeat improvements for [article]“
/lokal-seoLocal SEO”/lokal-seo for [business] in [city]“

Category 6: Business & Productivity (10 Commands)

CommandWhat It DoesExample Prompt
/swotSWOT analysis”/swot for [idea/company]“
/pitchElevator pitch”/pitch 60 seconds for [product]“
/meetingAgenda/minutes”/meeting agenda for sprint review”
/emailEmail draft”/email to [person], purpose, tone: professional”
/reportReport”/report from <data>[data]</data>”
/agendaMeeting agenda”/agenda 45 min on [topic]“
/todoPrioritize”/todo from <list>[tasks]</list>”
/brainstormIdeas”/brainstorm 15 ideas for [problem]“
/decisionDecision matrix”/decision A vs B by [criteria]“
/summarySummary”/summary of <text>[text]</text>, 5 points”

Claude tip: /summary and /analyze are Claude’s forte. Drop in full long documents and it processes 200k-token context windows.

Category 7: Data Analysis & Tables (10 Commands)

CommandWhat It DoesExample Prompt
/analyzeAnalyze”/analyze <data>[data]</data>”
/matrixMatrix”/matrix [options] by [criteria]“
/tableTable”/table [data] as markdown table”
/chartChart code”/chart Python code for this data”
/compareCompare”/compare [A] vs [B]“
/forecastForecast”/forecast trend from [data]“
/segmentSegment”/segment list by [criterion]“
/surveySurvey”/survey 10 questions on [topic]“
/insightInsights”/insight from <text>[report]</text>”
/csvAs CSV”/csv structure <text>[data]</text>”

Claude tip: /table and /matrix produce clean Artifacts in Claude’s interface, copyable and renderable directly in the UI.

Category 8: Social Media (10 Commands)

CommandWhat It DoesExample Prompt
/postCreate a post”/post LinkedIn about [topic]“
/threadCreate a thread”/thread X about [topic], 7 tweets”
/captionWrite a caption”/caption Instagram for [photo]“
/hashtagsGenerate hashtags”/hashtags 15 for [topic]“
/bioWrite profile bio”/bio for [person], 160 characters”
/replyWrite a reply”/reply to comment: [text]“
/viralCreate viral post”/viral PAS framework for [topic]“
/pollCreate a poll”/poll for [community]“
/carouselDesign carousel”/carousel 8 slides on [topic]“
/content-planPlan content”/content-plan 1 week for [channel]“

Category 9: Learning & Education (10 Commands)

CommandWhat It DoesExample Prompt
/explain-einfachExplain simply”/explain-einfach [concept]“
/quizCreate quiz”/quiz 10 questions on [topic] + answers”
/flashcardsMake flashcards”/flashcards for [topic]“
/lessonCreate lesson”/lesson [topic] for beginners”
/tutorAct as tutor”/tutor teach me [topic] with questions”
/analogyFind analogies”/analogy for [concept]“
/exampleProvide examples”/example 3 real-world examples for [topic]“
/study-planCreate study plan”/study-plan [topic] in 6 weeks”
/eli5Explain simply”/eli5 [topic]“
/fehlerCommon mistakes”/fehler common mistakes learning [topic]”

Claude tip: /tutor and /explain-einfach shine here. Claude explains with nuance and patience without oversimplifying to the point of inaccuracy.

Category 10: Prompt Engineering & Meta (10 Commands)

CommandWhat It DoesExample Prompt
/templateCreate prompt template”/template for recurring [task]“
/frameworkDefine thinking framework”/framework analyze with [model]“
/personaAssign a role”/persona handle as [role]“
/roleplaySimulate scenario”/roleplay negotiation conversation”
/chainThink step-by-step”/chain think step by step”
/iteriereImprove iteratively”/iteriere improve 3 times”
/critiqueSelf-critique”/critique your own answer”
/improve-promptRefine prompt”/improve-prompt my prompt: [text]“
/systemSet behavior”/system respond as [role]“
/formatSpecify format”/format as markdown table”

Claude-Specific Power Tips

  • Use XML tags: <document>, <code>, <example>. Claude parses them more precisely than free text.
  • Projects: Save frequently used codes as Project Instructions, they apply automatically.
  • Artifacts: /table, /document, /code output as separate artifacts, copy-ready and versionable.
  • Long context: Paste entire codebases, books, transcripts. /analyze becomes genuine analysis, not sampling.
  • Claude Code CLI: For coding workflows, the terminal version is more powerful: direct file access, test runs, Git integration.

Further Reading

Key Takeaways:

  • The same 100 codes work in Claude as in ChatGPT, plus XML tags as amplifiers.
  • Claude’s strengths: long context (200k tokens), Artifacts for output, Projects for saved rules, Claude Code for development.
  • Wrap input in <code>, <document>, <transcript> for more precise processing.
  • For coding: Claude Code CLI beats the web version.
  • Marketing and writing: Claude produces more nuanced, longer-form content.

FAQ

Are the codes the same as ChatGPT?

Yes, slash codes are universal prompt conventions. With Claude, XML tags act as amplifiers, and responses are often more nuanced and hold up better over time.

What do XML tags provide?

Claude is trained to respect XML structures. Tags like <code>, <document>, <data> mark input cleanly. Result: more precise parsing and less confusion between instructions and content.

What are Artifacts?

Separate windows for code, tables, documents. Claude places /table or /code output there, ready to copy and edit. More practical than inline text for productive work.

How long can input be?

Up to 200k tokens (roughly 150,000 words), entire books or repos. This makes /analyze and /summary particularly strong in Claude for large documents.

Claude Code vs. web interface?

For coding, the CLI (claude in Terminal) is stronger. It reads and writes files directly and runs tests. The web version works best for text, analysis, and general prompts.

Claude or ChatGPT for prompts?

Claude for long documents, nuanced writing, and coding with Claude Code. ChatGPT for broader tool integration (DALL-E, web browse) and if you use GPTs or custom instructions. The codes work in both.

Sources and Further Reading

Back to Blog
Share:

Related Posts