Guides programmatic SEO—creating large numbers of SEO-optimized pages automatically using templates and structured data, rather than writing each page manually. Classic “mail merge” pSEO (one rigid template + swapped variables) often produced low differentiation and thin-feeling URLs. With AI used responsibly on top of the same data spine, you can scale per-URL customization—intent-aligned copy, section depth, FAQs, tone, localization—while still following evidence blocks, data tiers, and QA (se
Fetches programmatic-seo from kostja94/marketing-skills and configures it for Cursor.
3
Select Cursor when prompted
The CLI shows a list of agents. Use arrow keys and space to select Cursor:
◆ Which agents do you want to install to?
│
│ ── Universal (.agents/skills) ────────────────
│ · Cline · Codex · Goose · Windsurf
│ ●Cursor(selected)
│ · Cursor · Aider · Continue
4
Verify installation
Confirm successful installation by checking the skill directory location:
.cursor/skills/programmatic-seo
Restart Cursor to activate programmatic-seo. Access via /programmatic-seo in your agent's command palette.
⚠
Security Notice
We perform automated surface-level scans (Gen AI Scanner, Socket, Snyk) during installation. These checks detect common vulnerabilities but do not guarantee complete security. Always review skill source code and verify the publisher's reputation before production use.
Skills execute code in your environment. Always review source, verify the publisher, and test in isolation before production.
Guides programmatic SEO—creating large numbers of SEO-optimized pages automatically using templates and structured data, rather than writing each page manually. Classic “mail merge” pSEO (one rigid template + swapped variables) often produced low differentiation and thin-feeling URLs. With AI used responsibly on top of the same data spine, you can scale per-URL customization—intent-aligned copy, section depth, FAQs, tone, localization—while still following evidence blocks, data tiers, and QA (see Data strength hierarchy and AI-assisted generation below).
When invoking: On first use, if helpful, open with 1–2 sentences on what this skill covers and why it matters, then provide the main output. On subsequent use or when the user asks to skip, go directly to the main output.
Project context: If .claude/project-context.md or .cursor/project-context.md exists, read product/ICP sections before proposing playbooks or page types.
Definition
Programmatic SEO = Building a single template and populating it with data from a database, API, or spreadsheet to generate hundreds or thousands of unique pages. Each page targets a long-tail keyword (e.g., "best SEO tool in [city]," "[App A] + [App B] integration").
Key differences from traditional SEO: Technical (SEOs + engineers); long-tail focus; data-driven (data quality = success); automation; built for scale.
Classic limits vs AI-enhanced differentiation
Era
What breaks
What helps
Rigid pSEO
One template, minimal variance → similar titles/bodies, weak E-E-A-T, “obvious mail merge”
Still needs unique evidence per URL and selective indexation
AI-enhanced pSEO
Same structured rows (facts, SKUs, metrics) drive the page, but models add per-URL narrative: intros, FAQ depth, persona angles, localization, internal-link suggestions—higher differentiation at scale
Facts stay in your data layer; AI shapes phrasing and structure, not invented numbers—see AI-assisted generation
Best-practice stance: AI is an accelerator and customizer, not a substitute for data defensibility (Tiers 1–5) or technical SEO (URLs, schema, CWV). Used well, it aligns with quality over quantity: fewer thin URLs, more distinct useful pages.
Structured information: locations, products, prices, features—must be accurate, complete, and add genuine value
Automation
Systems connecting data to templates; pages generated dynamically or published in bulk
AI layer (optional)
On grounded inputs (row JSON + rules), generates varied copy, FAQ expansions, and section emphasis per URL—reduces “same template” fatigue while staying auditable
Page Playbook Matrix (skills/pages)
Page types in this library live under pages/{brand|content|legal|marketing|utility}/. Use the matrix below to map search pattern → playbook → which *-page-generator skill to open for structure, copy, and schema—not every folder is a good fit for mass-generated URLs.
Playbook
Example intent / keyword pattern
Page skill (name)
Path (reference)
Alternatives / comparisons
"[Competitor] alternatives", "X vs Y"
alternatives-page-generator
pages/marketing/alternatives
Integrations
"[Product A] [Product B] integration"
integrations-page-generator
pages/marketing/integrations
Category / catalog
Faceted listings, product grids
category-page-generator, products-page-generator
pages/marketing/category-pages, products
Glossary / definitions
"what is [term]", term landings
glossary-page-generator
pages/content/glossary
FAQ / Q&A
Question banks, PAA-style pages
faq-page-generator
pages/content/faq
How-to / procedures
Step libraries, "[how to] [task]" blocks in templates
Usually not mass programmatic (single primary URL or compliance-heavy): pages/brand/* (home, about, contact), pages/legal/*, most pages/utility/* (404, status, signup-login, etc.)—treat as one-off or policy-driven, not template×data scale.
Data-driven content unique to each page (tables, lists, verified stats); differentiates from thin content
Decision
Comparison, recommendation, or next steps
FAQ
Frequently asked questions
CTA
Call-to-action
Evidence block = Real, structured data per page (business listings, pricing, reviews, verified stats). Ensures each page delivers genuine value, not recycled boilerplate with swapped variables.
Data strength hierarchy (defensibility)
Strongest programmatic pages are fueled by what only your product (or your customers inside your product) can produce—especially templates, exports, and generated artifacts. Third-party or scraped lists alone are the weakest foundation.
Tier
Source
Examples
Relative risk
1 — Product-generated
Assets created or rendered by your product: page/layout templates, email/Notion/code templates, export packs, generated previews, branded snippets, “built with [Product]” examples
Template gallery rows tied to real .json / CMS fields; screenshots of exports; unique preview URLs
Lowest when each URL shows distinct generated output
2 — Product-derived
Telemetry and in-product data you own: aggregates, cohorts, benchmarks, feature adoption
“Teams in [industry] median time-to-value” from your warehouse (aggregated)
Low if aggregated / anonymized and policy-compliant
3 — UGC / customer
Reviews, submissions, showcase items, moderated community content
Showcase grid; verified quotes
Medium—needs moderation + consent
4 — Licensed / partner
Exclusive feeds, co-marketing datasets
Partner pricing tiers; licensed industry stats
Medium—contract and citation discipline
5 — Public / scraped
Open web, directories, generic enrichment
Name/address fills; commodity facts
Highest—needs editorial layer, fact-checking, and a real Evidence block
Why Tier 1 (templates & generated content) wins: Pages built from your template system carry proprietary structure, variables, and brand-safe blocks—harder for competitors to copy verbatim and easier to prove uniqueness (embeds, downloads, IDs). Pair with template-page-generator when the UX is “browse gallery → use template.”
Tier 2 — Product-derived (practical)
What it is
What to watch
Metrics from your backend, data warehouse, or support/CRM exports: activation rates by segment, integration popularity, error budgets, time-to-value—not generic “industry reports.”
Privacy & ToS: Minimum cell sizes; no individual identification; document what was aggregated and over what window.
Good fit when you can show “only we could know this because it runs in our product.”
Stale data hurts trust: pipeline jobs, “as of [date]” labels, automated invalidation.
AI here: Use models to turn structured aggregates into prose (intro paragraphs, “what this means for [segment]”)—input must be verified numbers/tables from your pipeline, not free-form invention. Keep a machine-readable table or JSON on-page or in appendix so claims stay auditable.
Tier 3 — UGC / customer (practical)
What it is
What to watch
Quotes, reviews, showcase submissions, community templates—per-user artifacts with consent.
Moderation: spam, PII, competitor attacks; consent for name/logo use; schema (Review, CreativeWork) only when accurate.
Strong when combined with Tier 1 (e.g. “customer-built template” gallery).
Volume without quality → thin pages; cap or score submissions.
AI here: Summarize long reviews into bullets; generate draft alt text for images; cluster submissions into topic pages—always human approve before publish. Do not fabricate testimonials.
Contract scope: Which fields can appear on which URLs; attribution line; DMCA / trademark on logos.
Often one feed → many URLs; uniqueness must come from your framing, comparison logic, or calculator—not the raw feed alone.
Refresh cadence tied to partner SLAs.
AI here: Draft comparison copy and FAQs from a fixed attribute table (license + partner rules); never invent SKUs or prices—pull from feed, let AI phrase and shorten.
Tier 5 — Public / scraped (practical)
What it is
What to watch
Open data, directories, Wikipedia-style facts, enrichment of public entities.
Highest duplicate/thin risk: everyone has the same facts; you must add synthesis, editorial angle, or a unique tool (calculator, filter) on top.
Entity SEO and citations matter: link to authoritative sources; date-stamp volatile facts.
Plan for pruning or noindex on low-traffic thin URLs.
AI here: Use models to structure messy public text into tables, outline sections, suggest internal links—then fact-check names, numbers, and dates. Do not use AI to invent statistics or citations; treat output as draft until verified.
AI-assisted generation (cross-tier)
Why AI fits modern pSEO: Early programmatic SEO earned a bad reputation because templates were frozen and copy was interchangeable—little real differentiation per query. LLMs, when grounded on each row’s facts and your brand rules, make it practical to customize headlines, intros, FAQs, and “why this page matters” per URL without hand-writing thousands of pages. That moves execution closer to best practices (intent match, helpful content, unique value) at scale, provided you do not let the model invent data.
Principle
Why
Ground AI in structured inputs
Pass JSON/CSV rows (tier, source URL, metrics) into prompts; forbid hallucinated numbers in system prompts.
Separate “facts” from “phrasing”
Data layer = source of truth; AI = tone, shortening, localization, FAQs, per-segment emphasis—never the other way around.
Vary structure, not only adjectives
Ask for different section order, FAQ count, or “beginner vs power user” angles by intent flags in the row—reduces template sameness.
Human or automated QA
Spot-check high-traffic URLs; block publish if required fields empty or citation missing.
Disclose when useful
e.g. “Intro generated with AI; figures from [internal report, Q3 2025].” Builds trust and matches policy trends.
When AI generation is a strong lever: Tiers 2–5—where raw material is already tabular or repetitive but needs readable, differentiated copy at scale. Tier 1 still benefits from AI (drafts from export JSON), but the differentiator remains the product artifact itself.
Operational requirements (all tiers)
Requirement
Practice
Provenance
Log data sources; track origin per field
Freshness rules
e.g., ratings every 90 days, prices every 30 days, template version bumps when layouts change
Prefer 1–2 over 5
Fill evidence with product-generated or product-derived data before reaching for public scraping
AI governance
Structured inputs only; no unverified numbers; moderation on UGC; optional disclosure
Clean & merge
Deduplicate keys; drop rows that produce duplicate intents
Ideal Use Cases
For which page-generator skill to use, see Page Playbook Matrix above. Generic patterns:
Use case
Example
Location-specific pages
"Plumber in [city]," "Best restaurants in [neighborhood]" with real local data
Product comparison
"[Product A] vs [Product B]" with structured specs
Alternatives pages
"[Competitor] alternatives" at scale; 50+ competitors; see alternatives-page-generator
Each page must add meaningful page-specific content beyond simple data swaps
Evidence block
Tables, lists, examples with real numbers/attributes on every page
Semantic HTML
Proper structure; conditional logic to avoid empty or repetitive sections
Internal linking
Link related programmatic pages; compounds traffic and indexation
Technical Considerations
Topic
Practice
Subfolder vs subdomain
Prefer subfolders (yoursite.com/integration/slack-notion/) over subdomains (integrations.yoursite.com/...) so authority consolidates on the primary domain; see url-structure, domain-architecture if restructuring
Selective indexation
Don't index all pages; use noindex rules for lo
Implementation Guide
Prerequisites
›Claude Desktop or compatible AI client with skill support
›Clear understanding of task or problem to solve
›Willingness to iterate and refine outputs
Time Estimate
15-45 minutes depending on use case complexity
Steps
1Install skill using provided installation command
2Test with simple use case relevant to your work
3Evaluate output quality and relevance
4Iterate on prompts to improve results
5Integrate into regular workflow if valuable
Common Pitfalls
⚠Expecting perfect results without iteration
⚠Not providing enough context in prompts
⚠Using skill for tasks outside its intended scope
⚠Accepting outputs without review and validation
Best Practices
✓ Do
+Start with clear, specific prompts
+Provide relevant context and constraints
+Review and refine all outputs before using
+Iterate to improve output quality
+Document successful prompt patterns
✗ Don't
−Don't use without understanding skill limitations
−Don't skip validation of outputs
−Don't share sensitive information in prompts
−Don't expect skill to replace human judgment
💡 Pro Tips
★Be specific about desired format and style
★Ask for multiple options to choose from
★Request explanations to understand reasoning
★Combine AI efficiency with human expertise
When to Use This
✓ Use when
Use when skill capabilities match your task, clear ROI on time saved, and you can validate outputs. Best for repetitive tasks, learning, and quality improvement.
✗ Avoid when
Avoid when task requires deep expertise you can't validate, involves sensitive decisions, or when learning process is more valuable than speed of completion.
Learning Path
1Familiarize yourself with skill capabilities and limitations
2Start with low-risk, non-critical tasks
3Progress to more complex and valuable use cases
4Build expertise through regular use and experimentation
