Documentation
ICP Playbooks

Workflow Bundles

AgentSEO is most useful when customers do not have to assemble a random endpoint chain. These bundles package the API around the jobs a vibe marketer, SEO operator, or coding agent actually needs to finish.

Need first-call setup?

Start with quickstart, then come back here once the API key and job contract are working.

Need exact contracts?

Use the API reference for request bodies, response fields, examples, and credit notes.

Need product context?

Use cases explain where AgentSEO fits in agents, apps, agencies, and monitoring systems.

How to choose the right bundle

If the user asks what to publish next, start with the competitor gap or launch workflow.
If the user asks why traffic dropped, start with the refresh workflow.
If the user asks about ChatGPT, Perplexity, AI Overview, or citations, start with the AI visibility workflow.
If the user asks to scale many pages, start with the programmatic safety workflow.
If the user asks for backlinks or authority growth, start with the backlink opportunity workflow.
If the user is in Claude Code, return the workflow steps plus review gates, not just endpoint names.

Launch an SEO Page Safely

Move from keyword idea to publish-ready page without missing intent, schema, metadata, internal links, or conversion basics.

Credit posture: Mostly deterministic 2-credit steps after any upstream paid-data discovery. Keep the first pass narrow, then expand only the winner.

Who it is for

Founder-marketers, technical PMMs, and Claude Code users shipping one new SEO page.

When to use it

Use this when the page does not exist yet, or when a landing page is being rebuilt from scratch.

Endpoint chain

1/content/serp-outline

Create the page shape from the target keyword and SERP expectations.

2/content/keyword-map

Confirm whether the keyword deserves a new URL or should map to an existing page.

3/content/brief

Turn the outline into a writer-ready or agent-ready production brief.

4/content/title-meta

Generate search title and meta options that match the page promise.

5/content/schema-plan

Prepare structured-data recommendations and validation checks.

6/content/internal-links

Choose natural links from supplied related pages before publishing.

7/content/technical-qa

Run the final indexability, canonical, heading, metadata, schema, and link gate.

8/page/cro-qa

Check whether organic visitors can convert after the page earns traffic.

Success signals

  • One primary URL owns the keyword intent.
  • The page has publish-ready metadata, headings, schema guidance, and internal-link placements.
  • CTA, proof, and objections are handled before traffic arrives.

Risk controls

  • Do not publish automatically from an agent workflow.
  • Run technical QA after the final HTML or markdown exists, not before the real page shape is known.
  • Track the keyword/page with /rank/track after publishing so the loop measures outcomes.

Refresh a Declining or Stalled Page

Find whether the page needs a focused refresh, consolidation, differentiation, technical cleanup, or monitoring.

Credit posture: Start with the cheapest supplied-data checks when you already have exports. Use paid discovery only when the source of the problem is unclear.

Who it is for

SEO operators, agencies, and content teams maintaining pages that already have impressions or rankings.

When to use it

Use this when an existing URL has lost rank, stopped growing, or no longer matches the SERP.

Endpoint chain

1/domain/relevant-pages

Identify winners, decliners, recovery targets, and growth candidates from page-level signals.

2/domain/ranked-keywords

Pull the ranked keyword inventory for the page, subdomain, or domain when live search data is needed.

3/content/cannibalization

Check whether another owned URL is competing for the same query or intent.

4/content/refresh-brief

Create the focused edit plan: sections, proof, internal links, schema, and validation.

5/content/draft-qa

QA the edited draft before publishing so weak sections and proof gaps are caught.

6/content/technical-qa

Make sure the refreshed page is technically safe to re-ship.

7/serp/volatility

Use snapshots to decide whether movement is page-specific or SERP-wide.

Success signals

  • The workflow chooses refresh, consolidate, differentiate, technical fix, or monitor.
  • The edit brief is narrow enough for a writer or coding agent to execute.
  • Rank movement is interpreted against SERP volatility instead of treated as isolated failure.

Risk controls

  • Avoid rewriting the whole page when only a few intent gaps matter.
  • Do not redirect, canonicalize, or consolidate until cannibalization evidence is clear.
  • Keep before/after rank snapshots for the same keyword, URL, location, and device.

Weekly AI Visibility Review

Create a stable monitoring set, inspect AI surfaces, and turn visibility gaps into content or citation actions.

Credit posture: Prompt-set creation is deterministic. Live SERP or AI Overview checks should be limited to high-value prompts and reviewed weekly.

Who it is for

Vibe marketers, PMMs, and agency teams reporting on answer-layer visibility.

When to use it

Use this when the team needs a repeatable AI visibility routine instead of ad hoc prompt screenshots.

Endpoint chain

1/ai-visibility/prompt-set

Build the stable prompts, platforms, competitor checks, citation fields, and owned-asset mapping.

2/ai-overview/extract

Inspect whether Google AI Overview appears for the highest-value query samples.

3/llm-mentions/track

Track brand, competitor, citation, and first-mention movement when LLM monitoring is active.

4/serp/volatility

Separate answer-layer changes from broader SERP movement.

5/content/action-plan

Turn missing citations, weak owned assets, and competitive gaps into the next weekly work plan.

Success signals

  • Prompts are grouped by persona, funnel stage, topic, platform, and owned asset.
  • Reports separate mention, citation, first mention, competitor presence, and next action.
  • The output routes to content refresh, new content, citation research, or monitoring.

Risk controls

  • Do not compare random one-off prompts week to week.
  • Keep platform and prompt wording stable unless there is a deliberate measurement change.
  • Treat AI answers as sampled signals, not absolute rank-tracker truth.

Programmatic SEO Safety Check

Decide whether a programmatic page system is useful enough to scale before it creates indexation risk.

Credit posture: Deterministic planning and QA are cheap compared with scaling weak pages. Spend paid discovery only on sample templates and representative head terms.

Who it is for

Builders and growth teams creating repeatable page sets from templates, datasets, or CMS records.

When to use it

Use this before generating hundreds of pages or after a template exists but before broad indexation.

Endpoint chain

1/content/programmatic-template

Score template readiness, uniqueness, data quality, indexation gates, internal links, and launch QA.

2/content/keyword-map

Map keyword patterns to page types and catch missing or duplicated URL ownership.

3/content/cannibalization

Check whether generated pages would compete with each other or existing pages.

4/site/sitemap-audit

Confirm valuable pages are discoverable and noise pages are not polluting the sitemap.

5/content/technical-qa

QA representative rendered pages for metadata, canonicals, headings, schema, and links.

Success signals

  • The template has unique fields, useful public value, and safe indexation rules.
  • Generated URLs have a clear internal-link architecture and sitemap strategy.
  • The workflow can say do_not_scale_yet when pages are thin or duplicated.

Risk controls

  • Start with a small sample set before generating at scale.
  • Block indexation for low-data, duplicate, or near-empty pages.
  • Review representative rendered pages, not only the template config.

Competitor Content Gap Sprint

Turn competitor keywords, pages, and topics into a buyer-stage content matrix and execution plan.

Credit posture: Paid competitor discovery is useful for the shortlist. The matrix and action-plan steps work from supplied exports when you already have data.

Who it is for

Agencies, technical marketers, and founders choosing where to compete next.

When to use it

Use this when the team knows competitors matter but needs to decide what to create or refresh first.

Endpoint chain

1/domain/competitors

Find realistic search competitors before choosing who to compare against.

2/domain/intersection

Compare shared and competitor-exclusive keyword opportunities.

3/pages/intersection

Inspect exact page-level keyword gaps for the most important competitor pages.

4/content/competitor-gap-matrix

Group missing topics, weak coverage, page types, and buyer stages into a matrix.

5/content/action-plan

Turn the matrix into a capacity-aware 30-day roadmap.

6/content/brief

Generate the first production brief only after the gap is prioritized.

Success signals

  • The team can see page type, buyer stage, gap reason, impact, and effort.
  • The action plan separates new pages, refreshes, internal links, schema, and monitoring.
  • Briefs are generated for prioritized opportunities, not every possible keyword.

Risk controls

  • Do not chase every competitor-exclusive keyword.
  • Filter by business relevance and buyer stage before briefing.
  • Use competitor evidence to choose the page type, not to copy the competitor page.

Backlink Opportunity Sprint

Turn competitor link gaps, backlink prospects, and owned assets into safe outreach, reclaim, broken-link, resource-page, and digital PR actions.

Credit posture: Fresh backlink discovery uses paid backlink endpoints. /backlinks/opportunity-finder itself costs 2 credits because it works from supplied prospect data.

Who it is for

Agencies, founder-marketers, and technical marketers who need safe link acquisition priorities, not raw backlink exports.

When to use it

Use this after backlink competitor discovery, page intersection, a CSV prospect export, or a manual list of relevant source pages.

Endpoint chain

1/backlinks/competitors

Find backlink-profile competitors when you do not know whose link graph to inspect first.

2/backlinks/page-intersection

Find source pages that link to competitors but not to your domain.

3/backlinks/list

Inspect exact backlink rows for selected domains, lost links, broken targets, or outreach references.

4/backlinks/opportunity-finder

Prioritize safe outreach, reclaim, broken-link, resource-page, and digital PR opportunities from supplied prospects.

5/content/technical-qa

Validate the destination asset before pitching it as a link target.

6/content/action-plan

Schedule outreach, asset refreshes, and reclaim work into a practical weekly plan.

Success signals

  • Each prospect has a score, opportunity type, recommended owned asset, outreach angle, and risk flags.
  • Spammy, paid-link, weak-relevance, or no-asset prospects are flagged before outreach.
  • The plan separates link-gap outreach from reclaim, broken-link, resource-page, and digital PR work.

Risk controls

  • Manually qualify every source before outreach.
  • Do not buy links or automate spam outreach.
  • Pitch useful owned assets or original information, not generic homepage links.

Next: run one bundle in the playground

Start with one real ICP job, keep the first request narrow, and expand only when the output points to a higher-confidence next step.

Open playground