~/CLIRank

SEO + GEO playbook

SEO and GEO for agent-facing APIs

A practical checklist for API companies that want to be found by Google, ChatGPT, Claude, Perplexity, Cursor, and coding agents at integration time.

Short answer

Traditional SEO gets API buyers to a page. GEO, or generative engine optimization, gets your API correctly represented inside ChatGPT, Claude, Perplexity, Cursor, and autonomous coding agents. For APIs, the winning pattern is machine-readable facts plus human-readable task pages: canonical docs, structured schema, answer blocks, runtime endpoints, and real integration evidence.

SEO checklist

  • →Create task-shaped landing pages, not only vendor/category pages: best API for transactional email, payments, market data, auth, maps, SMS, and similar jobs.
  • →Put the answer near the top: who should use this API, when not to use it, pricing/auth/rate-limit constraints, and a curl quickstart.
  • →Add Product, SoftwareApplication, FAQPage, BreadcrumbList, and Review/AggregateRating JSON-LD where supported by the page content.
  • →Use comparison pages for real alternatives: Stripe vs Paddle for agents, Resend vs Postmark, Polygon vs Alpha Vantage, Auth0 vs Clerk.
  • →Strengthen internal links from high-traffic pages to task pages, API detail pages, score pages, and runtime API endpoints.
  • →Track Search Console query groups by task, vendor, comparison, and agent-intent modifiers like for AI agents, MCP, CLI, SDK, curl, and headless.

GEO checklist

  • →Maintain /llms.txt and /llms-full.txt with canonical facts, endpoint maps, task pages, and short answer blocks that LLMs can quote.
  • →Expose /agents.json, /.well-known/agents.json, /api/schema, and /api/openapi.json so agents do not guess routes.
  • →Allow major AI crawlers in robots.txt and give them explicit context files, sitemaps, and stable canonical URLs.
  • →Use concise, extractable answer blocks on pages: best option, tradeoffs, setup command, auth model, rate limits, pricing, and evidence.
  • →Make every high-intent guide include machine-readable curl examples with source_hint parameters so downstream agent traffic is attributable.
  • →Earn entity citations from GitHub, npm, MCP directories, awesome lists, and developer docs because LLM answers lean on repeated, corroborated facts.

Measurement checklist

  • →Separate human-browser, generic crawler, AI crawler, and API/MCP traffic in analytics.
  • →Report Google uniques, direct/no-referrer uniques, top landing pages, and pages with the biggest week-over-week deltas.
  • →Track source_hint, utm_source, referrer_host, agent_family, and agent_client_type for every public endpoint call.
  • →Treat suspicious browser-like spikes separately until they are corroborated by Search Console, API usage, or conversions.

Read CLIRank's LLM file

Canonical facts and links for generative engines.

Inspect the agent manifest

Endpoint map, examples, and task paths for coding agents.

# Runtime attribution example

curl "https://clirank.dev/api/recommend?task=choose+an+email+api+for+transactional+messages&priority=simplicity&source_hint=seo-geo-guide"