Prompts & Workflows
Overview
Tools answer one question at a time. Prompts run a workflow. Each prompt is a vetted recipe — "audit this site," "draft a content strategy," "write a monthly report" — that calls 5–15 tools in the right order, gathers the data, and gives the AI assistant a structured frame to write from.
Use a prompt instead of a tool when you want the model to do the whole job rather than answer a single question. Most clients live inside seo-audit and monthly-seo-report; the rest are situational.
How prompts work
- Your AI client (Claude Desktop, ChatGPT, Cursor) lists available prompts the same way it lists tools.
- Pick a prompt, fill in the arguments (usually
domain,site_id, optional date ranges or competitors), and the prompt runs. - Internally the prompt calls TransformSEO MCP tools, collates results, and hands the model a structured plan plus the data needed to write a report or recommendation.
- Free-tier teams get the
monthly-seo-reportprompt only. Starter unlocks the workflow set; Pro/Agency get the full catalog.
seo-audit
Comprehensive SEO audit covering technical health, on-page optimization, content quality, and backlink profile. Calls get-site-overview, get-page-audit, get-lighthouse-audit, get-backlink-summary, get-anchor-text-analysis, and the on-page crawl readers when a fresh crawl is available. Output is a prioritized issue list and an action plan.
Run theseo-auditprompt withdomain=elysiumpools.comandsite_id=1to generate the quarterly review.
monthly-seo-report
The deliverable agencies actually send to clients. Pulls month-over-month deltas from get-search-console-data, get-analytics-data, get-keyword-ranking-history, get-backlink-history, and get-llm-mention-history, then writes a narrative summary with wins, losses, and what to do next month. Available on every plan including free.
content-strategy
Builds a content strategy from competitor research and keyword opportunity. Uses get-competitor-domains, get-domain-intersection, get-keyword-ideas, get-keyword-volume, get-keyword-difficulty, and get-search-intent to identify gap topics, cluster them, and recommend a publishing cadence.
copywriting
Page-level copywriting brief grounded in real SERP data. Pulls top-ranking pages via get-serp-results, parses their content with get-content-parsing, and writes a brief that tells the writer what intent to match, what entities to cover, and where existing pages fall short.
schema-markup
Generates the JSON-LD schema for a page. Inspects the live page with get-content-parsing + get-page-audit to pick the right schema type (Article, Product, LocalBusiness, FAQ, HowTo, etc.) and emit valid markup with all required and recommended properties populated from the actual content.
site-architecture
Maps and re-plans a site's information architecture. Uses submit-onpage-crawl → get-onpage-crawl-pages → get-onpage-links to model the existing structure, identifies orphan pages and weak topical clusters, and proposes a new IA + internal linking plan.
ai-seo
AI search visibility audit. Calls get-llm-mentions, get-llm-mention-top-domains, get-llm-mention-top-pages, and get-aeo-health-report to show who ChatGPT, Claude, Gemini, and Perplexity cite when they answer for the brand's keywords — and what to fix to start being one of them.
competitor-analysis
Side-by-side competitor teardown. Pulls get-competitor-domains, get-domain-intersection, get-page-intersection, get-backlink-competitors, and get-anchor-text-analysis to compare visibility, content, links, and anchor profiles. Output highlights specific keywords and pages where the competitor is winning and what's defensible to take.
programmatic-seo
Plans a templated landing-page program. Validates seed lists with get-keyword-volume and get-keyword-difficulty, checks SERP types with get-serp-results to confirm intent matches a templated page, and outputs the page count, URL pattern, and content variables needed.
page-cro
Conversion-rate review of a specific URL. Combines get-page-audit, get-lighthouse-audit, and get-content-parsing with conversion heuristics (above-the-fold clarity, value prop, CTA contrast, social proof, friction) to produce a prioritized fix list.
ab-test-setup
Designs an A/B test for a page. Uses get-analytics-data to get baseline traffic, calculates the sample size and runtime needed for statistical significance, and proposes a hypothesis grounded in the page's specific weaknesses.
analytics-tracking
Tracking-plan generator. Inspects the site with get-content-parsing to identify conversion events worth tracking and emits a structured event taxonomy for GA4 / GTM with parameter definitions.
Tips
- Prompts make great onboarding deliverables — run
seo-auditon day one of a new client engagement to set baseline. - The token cost of a prompt run is paid by your AI assistant's provider, not by TransformSEO credits. Tool calls inside the prompt do consume credits.
- If your AI client doesn't expose prompts, you can still ask for the same workflow in plain English — the model will compose the tool calls itself; it just takes more turns.