# Stephen Sumner

> British-Swedish international SEO consultant and one of the most experienced SEO specialists working in Sweden. 27+ years in search: local SEO, the legal sector, and AI search visibility. Founder of Local SEO Ltd. Based in Gothenburg, Sweden, working worldwide.

## Pages
- [Home](https://stephensumner.com/): Projects, experience and an overview of Stephen Sumner's work in search.
- [Blog](https://stephensumner.com/blog/): Long-form writing on AI search, local SEO and the legal sector.
- [llms.txt v2: what changed and what to do about it](https://stephensumner.com/blog/llms-txt-v2/): The August 2026 revision of the llms.txt proposal explained — link relations (rel="alternate" type="text/markdown", rel="describedby"), two valid markdown URL forms, defined subpath scope, the retirement of llms_txt2ctx and the Optional section's mechanical meaning — with an honest read on measurable impact (near zero for most sites, but an hour's work), a six-step checklist and a copy-paste template.
- [The directory moat: data, audience and AI](https://stephensumner.com/blog/business-directory-moat/): A thesis on why business directories are a durable asset in the AI-search era — owned, structured business data; an audience you reach directly rather than renting from Google or Meta; and a clean, citable source that AI answer engines quote. Explains the local-SEO foundation (citations, NAP consistency, vertical authority), why directories are well-shaped for AI citation, and the flywheel between the three. Written as the thesis behind acquiring JusticeDirectory.co.uk, a UK legal-sector directory used to generate leads for Stephen Sumner and partner companies.
- [The five pillars of eCommerce SEO, rebuilt for the AI-search era](https://stephensumner.com/blog/ecommerce-seo-four-pillars/): A 2026 rebuild of a long-standing framework — technical, authority, relevance and experience, plus a new fifth pillar for AI search — covering crawlability and schema, E-E-A-T and earned links, search intent, Core Web Vitals (LCP, INP, CLS) and getting cited by AI systems.
- [The 120-Point AI Readiness Framework](https://stephensumner.com/blog/ai-readiness-framework/): A six-category rubric for auditing whether a site is eligible to be cited by AI search, what its content is worth citing, and whether it can be extracted cleanly. Updated September 2026 with ten gap-diagnosis characteristics (adding Fresh), freshness windows in ChatGPT and Gemini retrieval, the pre-search brand shortlist, HTTP 499 retrieval eligibility, llms.txt as a supply-chain attack surface, and where AI influence appears in analytics. Updated September 2026 with the unscored agent discovery ladder (Cloudflare's five-level model), graduated WebMCP and Site-as-MCP, and the llms.txt v2 link-relation requirement.
- [How I built this site: near-perfect PageSpeed, no framework](https://stephensumner.com/blog/how-this-site-is-built/): A practical how-to behind the site's PageSpeed scores — a clean 100 for SEO, accessibility and best practices from hand-written static HTML, self-hosted fonts, sized webp images, and one _headers file deployed free to Cloudflare's edge. Updated September 2026: a git-based deploy with a one-script build, axis-restricted variable fonts with metric-matched fallbacks, cross-document view transitions, a CSS scroll-driven progress bar, 103 Early Hints, prerender via the Speculation-Rules header, markdown twins with content negotiation, RSS/JSON feeds and a robots.txt Content-Signal.
- [About](https://stephensumner.com/about/): Background, focus areas (AI search, legal sector, local SEO, systems) and career.
- [Where to find me](https://stephensumner.com/where-to-find-me/): The verified, canonical list of Stephen Sumner's profiles across the web, anchored to Wikidata entity Q137536965.

## Contact
- Email: hello@stephensumner.com

## Feeds
- [RSS](https://stephensumner.com/feed.xml): every post, newest first.
- [JSON Feed](https://stephensumner.com/feed.json): the same, as JSON; each item carries its markdown twin URL.

## MCP
- [MCP server](https://mcp.stephensumner.com/mcp): every post and page as tools and resources (list, read, search) over the Model Context Protocol, Streamable HTTP, no auth.
- [API catalog](https://stephensumner.com/.well-known/api-catalog): RFC 9727 linkset of every machine interface (MCP, content index, feeds).

## Agent skills
- [Skills index](https://stephensumner.com/.well-known/agent-skills/index.json): the AI readiness audit, llms.txt v2, the static-site recipe and the eCommerce pillars as SKILL.md files an agent can load.

## Full text
- [llms-full.txt](https://stephensumner.com/llms-full.txt): every post in full, as one Markdown file, newest first.