platform.claude.com — llms.txt & GEO score

Anthropic Developer Documentation

6 of 11 checks pass. What is left: enough links carry a description, descriptions say something concrete, descriptions add to the title, link titles are descriptive and skippable links marked Optional.

• Enough links carry a description 115 of 629 described (18%) — for a map-shaped file the bar is 60%, not every line
  3/10
• Descriptions say something concrete Average 25 characters across 115 described links
  2/7
• Descriptions add to the title 9 descriptions just repeat the link title
  3/4
• Points at clean markdown copies 629 of 629 end in .md or .txt (100%, detected by file extension)
  10/10
• Link titles are descriptive 31 weak titles (bare URLs, “here”, one-word labels)
  3/4
• A workable number of links 629 links, nearly all pointing at machine-readable copies — an exhaustive map rather than a short guide, which is a legitimate shape for a documentation site
  4/4
• Skippable links marked Optional Nothing is marked as skippable
  0/3
• No duplicate links All 629 URLs are distinct
  2/2
• List items are markdown links 12 of 641 list items are plain text, not links
  4/4
• Grouped so the map stays navigable Median heading group holds 19 links across 12 groups (12 sub-headings)
  3/3
• Links are absolute URLs All 629 links are absolute URLs
  1/1

5 of 7 checks pass. What is left: summary directly under the title and grouped into balanced sections.

• Exactly one H1 project title “Anthropic Developer Documentation”
  3/3
• Summary directly under the title An opening paragraph is there, but not as a "> " blockquote — clients that look for the quoted summary will miss it
  1/3
• Summary is a usable length 90 characters (from the opening paragraph)
  3/3
• Concrete language, not marketing filler No marketing filler detected
  3/3
• Grouped into balanced sections 3 H2 sections and 12 sub-headings, largest group holds 377 of 629 links
  3/4
• Conforms to the published grammar Markdown only, headings used as the spec defines them
  2/2
• Section names say what is inside All 3 headings are descriptive
  2/2

All 5 checks pass in this group.

• At the root of the domain you typed Found at the standard /llms.txt path
  2/2
• Sent as plain text Content-Type: text/plain; charset=utf-8
  1/1
• robots.txt exists 138 B of crawler rules
  1/1
• AI crawlers are not blocked No AI user-agent is disallowed at the root
  6/6
• robots.txt points at a sitemap A Sitemap: line is declared
  1/1

3 of 6 checks pass. What is left: homepage title is the right length, homepage has a meta description and canonical URL declared.

• Homepage title is the right length 15 characters, from title — shorter than search engines prefer
  1/2
• Title is not padded or repetitive No repeated terms
  1/1
• Homepage has a meta description 28 characters, from meta description — thin for a snippet
  1/2
• Canonical URL declared No rel="canonical" link
  0/1
• Language declared lang="en-US"
  1/1
• Social card image og:image is set
  1/1

4 of 5 checks pass. What is left: the index points at the full file.

• Publishes llms-full.txt Companion full-text file found
  3/3
• The index points at the full file Not linked from llms.txt
  0/1
• Index is a sensible size 66 KB
  2/2
• Server reports when it last changed Server sends Last-Modified
  1/1
• The index is kept current Last changed 0 days ago
  2/2

• Enough links carry a description Content & links +7

  Right now: 115 of 629 described (18%) — for a map-shaped file the bar is 60%, not every line

  Add ": " and one short line after every link. An undescribed link is a coin flip — the model has to open it to find out what it is.

  Where: Each "- [Title](url)" line

  1. Add ": " and one short line after every link — not most of them, every one.
  2. Describe what the page answers, not what it is called.
  3. An undescribed link is a coin flip: the model has to open it to find out.

  In your file

  line 34

  - [Documentation](https://platform.claude.com/docs/en/home.md)

  - [Documentation](https://platform.claude.com/docs/en/home.md): {one line on what this page covers}

  line 39

  - [Features overview](https://platform.claude.com/docs/en/build-with-claude/overview.md)

  - [Features overview](https://platform.claude.com/docs/en/build-with-claude/overview.md): {one line on what this page covers}

  line 42

  - [Intro to Claude](https://platform.claude.com/docs/en/intro.md)

  - [Intro to Claude](https://platform.claude.com/docs/en/intro.md): {one line on what this page covers}

  Placeholders in { } are yours to write — we do not invent text about your pages.

• Descriptions say something concrete Content & links +5

  Right now: Average 25 characters across 115 described links

  Write 60+ characters that add information the title does not: numbers, limits, formats, names.

  Where: The text after ": " on each link

  1. Write 60 characters or more, and make them add information the title does not.
  2. Cover every link: this is measured per LINK, not per description, so an undescribed line counts as zero.
  3. Name the concrete things on the page: numbers, limits, formats, model names.
  4. If the description could be pasted under any other link, it is not a description yet.

  In your file

  line 38

  - [Overview](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview.md) - Agent Skills

  - [Overview](https: {what is on the page — numbers, limits, formats}

  line 40

  - [Quickstart](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/quickstart.md) - Get started with Agent Skills in the API

  - [Quickstart](https: {what is on the page — numbers, limits, formats}

  line 41

  - [Quickstart](https://platform.claude.com/docs/en/get-started.md) - Get started with Claude

  - [Quickstart](https: {what is on the page — numbers, limits, formats}

  Placeholders in { } are yours to write — we do not invent text about your pages.

• Skippable links marked Optional Content & links +3

  Right now: Nothing is marked as skippable

  Move the nice-to-haves under ## Optional so a model on a tight budget knows what to drop first.

  Where: A new "## Optional" section at the end of llms.txt

  1. Move the nice-to-have links — history, press, careers — under "## Optional".
  2. By convention a client may skip that whole section when context is tight.
  3. This is you deciding what gets dropped first, instead of the model guessing.

  Do this

  ## Optional
  
  - [Company history](https://acme.example/about.md): Background, not needed to answer product questions

• Summary directly under the title Structure +2

  Right now: An opening paragraph is there, but not as a "> " blockquote — clients that look for the quoted summary will miss it

  Put the opening line in a > blockquote. The spec calls it optional, but it is the first thing a model reads and often the only thing it quotes back.

  Where: The line directly under the H1

  1. Add a single line starting with "> ".
  2. Answer three things in it: what you do, for whom, and what makes it specific.
  3. This is the sentence a model is most likely to quote back, so write it last, after the rest of the file.

  Instead of

  > We are an innovative company delivering world-class solutions.

  Do this

  > Industrial spare parts for food processing lines, shipped across Europe with same-week dispatch.

• Canonical URL declared Search-engine basics +1

  Right now: No rel="canonical" link

  Add <link rel="canonical"> so www and non-www are not counted as two competing sites.

  Where: The <head> of every page

  1. Add <link rel="canonical" href="..."> with the absolute, preferred URL.
  2. Pick one host — www or non-www — and make every canonical point at it.

  Do this

  <link rel="canonical" href="https://acme.example/">

• Descriptions add to the title Content & links +1

  Right now: 9 descriptions just repeat the link title

  A description that echoes the title costs context and adds nothing. Say what is on the page, not what it is called.

  Where: The text after ": " on each link

  1. Find the descriptions that just repeat the link title and rewrite them.
  2. Say what is ON the page — the numbers, limits, formats or names a reader would come for.
  3. If the description and the title carry the same information, one of them is wasted context.

  In your file

  line 101

  - [Streaming Messages](https://platform.claude.com/docs/en/build-with-claude/streaming.md) - Streaming messages

  - [Streaming Messages](https: {what is on the page — numbers, limits, formats}

  line 111

  - [Tool Runner (SDK)](https://platform.claude.com/docs/en/agents-and-tools/tool-use/tool-runner.md) - Tool runner (SDK)

  - [Tool Runner (SDK)](https: {what is on the page — numbers, limits, formats}

  line 275

  - [C#](https://platform.claude.com/docs/en/cli-sdks-libraries/sdks/csharp.md) - C# SDK

  - [C#](https: {what is on the page — numbers, limits, formats}

  Placeholders in { } are yours to write — we do not invent text about your pages.

• Grouped into balanced sections Structure +1

  Right now: 3 H2 sections and 12 sub-headings, largest group holds 377 of 629 links

  Group links under headings a reader would look for, and spread them — one group holding most of the file is the same as no grouping.

  Where: The "## " headings in llms.txt

  1. Split the link list into at least five "## " groups by what a reader would be looking for.
  2. Typical split: Docs, API, Guides, Product, Company, Optional.
  3. Grouping is what lets a model skip nine tenths of the file and still find the right page.

  Instead of

  ## Links
  
  - [A](...)
  - [B](...)
  - [C](...)

  Do this

  ## Docs
  
  - [A](...)
  
  ## API
  
  - [B](...)
  
  ## Optional
  
  - [C](...)

• Homepage has a meta description Search-engine basics +1

  Right now: 28 characters, from meta description — thin for a snippet

  Add a 70–160 character meta description. It is the snippet under your result, and assistants reuse it verbatim.

  Where: The <meta name="description"> tag on your homepage

  1. Target 70–160 characters; beyond that Google truncates it mid-sentence.
  2. Write it as the sentence you would want quoted, because assistants reuse it verbatim.

  Instead of

  <meta name="description" content="Welcome to our website.">

  Do this

  <meta name="description" content="Industrial spare parts for food processing lines, shipped across Europe with same-week dispatch and two-year warranty.">

• Homepage title is the right length Search-engine basics +1

  Right now: 15 characters, from title — shorter than search engines prefer

  Write a 30–60 character title that names the product and what it does.

  Where: The <title> tag on your homepage

  1. Target 30–60 characters; past 60 the result gets cut off in search.
  2. Put what you do before the brand name — the first words are what survives truncation.

  Instead of

  <title>Acme Parts | Acme | Industrial Spare Parts Supplier Europe Acme</title>

  Do this

  <title>Industrial spare parts for food lines — Acme</title>

• Link titles are descriptive Content & links +1

  Right now: 31 weak titles (bare URLs, “here”, one-word labels)

  Replace bare URLs and “click here” with the page’s real name — the title is half the signal.

  Where: The "[Title]" part of each link

  1. Replace bare URLs, "here", "click here" and one-word labels with the page’s real name.
  2. The title is half the signal — a model reads it before it reads your description.

  In your file

  line 69

  - [Effort](https://platform.claude.com/docs/en/build-with-claude/effort.md)

  - [{the page’s real name}](https://platform.claude.com/docs/en/build-with-claude/effort.md)

  line 136

  - [Dreams](https://platform.claude.com/docs/en/managed-agents/dreams.md)

  - [{the page’s real name}](https://platform.claude.com/docs/en/managed-agents/dreams.md)

  line 151

  - [Tools](https://platform.claude.com/docs/en/managed-agents/tools.md)

  - [{the page’s real name}](https://platform.claude.com/docs/en/managed-agents/tools.md)

  Placeholders in { } are yours to write — we do not invent text about your pages.

• The index points at the full file Depth & upkeep +1

  Right now: Not linked from llms.txt

  Mention llms-full.txt inside llms.txt. Guessing the filename works often enough to be tempting and not often enough to rely on.

  Where: Inside llms.txt

  1. Mention the full file in llms.txt itself, as a link or a line of context.
  2. Guessing the filename works often enough to be tempting and not often enough to rely on.

  Do this

  Full inlined text: https://acme.example/llms-full.txt

This score is Devsly’s own heuristic, not an official standard — llms.txt has no scoring body. It measures how well a file serves an AI assistant, and it does not predict whether any assistant will cite the site. The reading is a snapshot from 21 September 2026; run it again for the current state, or check a different site.