{
  "name": "sight4agents",
  "tagline": "Sites with sight.",
  "description": "Websites and applications built for humans and AI agents",
  "url": "https://sight4agents.com/",
  "brand": {
    "guide": "https://sight4agents.com/brand",
    "logo": "https://sight4agents.com/brand/logo.svg",
    "mark": "https://sight4agents.com/brand/mark.svg",
    "colors": {
      "ink": "#0E1020",
      "paper": "#F6F5F0",
      "iris": "#5B3DF5",
      "irisLight": "#A493FF",
      "signal": "#C6F432"
    },
    "fonts": {
      "display": "Space Grotesk",
      "body": "Inter",
      "mono": "JetBrains Mono"
    }
  },
  "version": "0.2.0",
  "last_updated": "2026-11-12",
  "contact": "hello@sight4agents.com",
  "endpoints": {
    "llms_txt": "https://sight4agents.com/llms.txt",
    "llms_full_txt": "https://sight4agents.com/llms-full.txt",
    "sitemap": "https://sight4agents.com/sitemap.xml",
    "search_json": "https://sight4agents.com/search.json?q={query}",
    "search_html": "https://sight4agents.com/search?q={query}",
    "mcp": "https://sight4agents.com/mcp",
    "markdown": "https://sight4agents.com/{page}.md"
  },
  "pages": [
    {
      "path": "/",
      "url": "https://sight4agents.com/",
      "markdown_url": "https://sight4agents.com/index.md",
      "title": "sight4agents: websites and apps built for humans and AI agents",
      "description": "sight4agents creates optimized websites and applications with native MCP (Model Context Protocol) support and human-in-the-loop change management, so agents can read and propose changes while people stay in control of what goes live.",
      "last_updated": "2026-11-12",
      "sections": [
        {
          "id": "overview",
          "heading": "Websites agents can ==see==, and humans sign off on.",
          "url": "https://sight4agents.com/",
          "text": "Websites agents can see, and humans sign off on.\nsight4agents creates optimized websites and applications with native Model Context Protocol (MCP) support and human-in-the-loop change management. Agents can navigate your content, search it, and propose changes. People review and approve every change before it goes live.\nGet early access\nHow it works"
        },
        {
          "id": "what-it-is",
          "heading": "What sight4agents is",
          "url": "https://sight4agents.com/#what-it-is",
          "text": "What sight4agents is\nMore and more of the web is read by software agents as well as by people. Most sites are still built only for browsers, so agents have to scrape layouts, guess at structure, and work around scripts.\nsight4agents builds sites and applications that serve both audiences from the same content: a fast, accessible experience for people, plus structured, machine-readable access for agents through MCP, Markdown, and JSON. When agents help maintain a site, every change they propose goes through a review workflow, so nothing ships without a human saying yes."
        },
        {
          "id": "how-it-works",
          "heading": "How it works",
          "url": "https://sight4agents.com/#how-it-works",
          "text": "How it works\nBuild: Your site or app is generated with semantic HTML, structured data, and clean, fast pages that work for people on any device.\nExpose through MCP: Each site can include a native MCP server, so AI agents can list, read, and search its content through tools instead of scraping pages.\nAgents propose: Agents can draft updates such as new copy, fixes, or new pages. Each one is submitted as a change proposal, never as a direct edit.\nHumans approve: A person reviews each proposal with a clear diff and preview, then approves, requests changes, or rejects it.\nShip with a record: Approved changes are published, and every step is recorded so you can see who proposed what, who approved it, and roll back if needed.\nRead the full workflow"
        },
        {
          "id": "features",
          "heading": "Features",
          "url": "https://sight4agents.com/#features",
          "text": "Features\nNative MCP: A Model Context Protocol endpoint ships with the site, with tools that let agents list pages, fetch content, and search.\nHuman-in-the-loop changes: Agents propose and people approve. Proposals carry diffs, previews, and a reason, so reviews are quick and informed.\nMachine-readable content: Every page is also available as Markdown, with an llms.txt overview, a JSON content index, a sitemap, and structured data.\nSemantic, accessible HTML: Proper landmarks, a clear heading hierarchy, and readable typography for screen readers, keyboards, and agents alike.\nFast and lightweight: Server-rendered pages with minimal dependencies and no client-side framework required to read the content.\nAudit trail and rollback: A history of every proposed, approved, and published change, so you can trace a decision or undo one."
        },
        {
          "id": "faq",
          "heading": "Frequently asked questions",
          "url": "https://sight4agents.com/#faq",
          "text": "Frequently asked questions\nWhat is sight4agents?\nsight4agents is a project that creates optimized websites and applications that serve both people and AI agents. It combines native MCP support with a human-in-the-loop change management workflow.\nWhy the name?\nIt is a play on words: a site with sight, for agents. Most websites are built only for human eyes. sight4agents builds sites that agents can actually see, through structured content and MCP, while humans keep the final say.\nWhat is MCP?\nThe Model Context Protocol (MCP) is an open protocol that lets AI applications connect to external tools and data through a standard interface. A site with an MCP server lets an agent call tools such as \"search\" or \"get page\" instead of scraping HTML.\nWhat does human-in-the-loop change management mean?\nAgents can suggest changes to a site, but they cannot publish on their own. Each change becomes a proposal that a person reviews and approves or rejects before anything goes live.\nIs the site still good for human visitors?\nYes. Pages are responsive, accessible, and fast. The agent-facing formats are generated from the same content, so people and agents always see the same information.\nHow can an agent read this website?\nStart with /llms.txt, fetch any page as Markdown by adding .md to its path, use /api/content.json or /search.json, or connect to the read-only MCP endpoint at /mcp. See For agents for details.\nHow do I get access?\nsight4agents is in early development. Use the early access link on this page to get in touch."
        },
        {
          "id": "early-access",
          "heading": "Get early access",
          "url": "https://sight4agents.com/#early-access",
          "text": "Get early access\nsight4agents is in early development. If you would like to build a site that works for both people and agents, send a short note about your project.\nEmail us for early access"
        }
      ]
    },
    {
      "path": "/how-it-works",
      "url": "https://sight4agents.com/how-it-works",
      "markdown_url": "https://sight4agents.com/how-it-works.md",
      "title": "How sight4agents works: build, MCP, and human-approved changes",
      "description": "How sight4agents builds agent-optimized websites, exposes them through native MCP, and routes every agent-proposed change through human review and approval before it goes live.",
      "last_updated": "2026-11-12",
      "sections": [
        {
          "id": "overview",
          "heading": "How it works",
          "url": "https://sight4agents.com/how-it-works",
          "text": "How it works\nsight4agents covers the whole lifecycle of a site: building it for people and agents, exposing it through MCP, and managing changes so agents can help while humans decide what goes live."
        },
        {
          "id": "build",
          "heading": "1. Build for people and agents",
          "url": "https://sight4agents.com/how-it-works#build",
          "text": "Build for people and agents\nSites start from structured content. That content is rendered into fast, semantic, accessible HTML for people, and into Markdown, JSON, and structured data for agents.\nSemantic HTML5 landmarks and a logical heading hierarchy\nJSON-LD structured data (Schema.org) on every page\nMarkdown versions of every page, plus llms.txt and llms-full.txt\nA sitemap, a JSON content index, and server-side search\nResponsive layouts with light and dark themes"
        },
        {
          "id": "mcp-native",
          "heading": "2. MCP-native by default",
          "url": "https://sight4agents.com/how-it-works#mcp-native",
          "text": "MCP-native by default\nEach site can expose a Model Context Protocol server. Agents connect over Streamable HTTP and use tools to work with the site directly: listing pages, reading content, and searching.\nRead-only tools are open to anyone. Tools that change content are designed to create proposals for review, not to publish directly."
        },
        {
          "id": "change-management",
          "heading": "3. Human-in-the-loop change management",
          "url": "https://sight4agents.com/how-it-works#change-management",
          "text": "Human-in-the-loop change management\nAgents are useful editors. They can fix typos, update outdated details, and draft new pages. But publishing is a decision for people. In sight4agents, every agent edit follows the same path:\nPropose: An agent submits a change proposal with the edited content and a short explanation of why.\nPreview: The proposal is rendered as a diff and a preview, so reviewers see exactly what will change.\nReview: A person approves it, asks for revisions, or rejects it. Nothing is live yet.\nPublish: Only approved changes are published.\nRecord: Every step is logged so changes can be traced and rolled back."
        },
        {
          "id": "roles",
          "heading": "Who does what",
          "url": "https://sight4agents.com/how-it-works#roles",
          "text": "Who does what\nAgents: Read and search content, then draft and submit change proposals with reasons.\nReviewers: Inspect diffs and previews, then approve, request changes, or reject.\nOwners: Decide who can review and publish, and can see the full history of every change."
        },
        {
          "id": "early-access",
          "heading": "Want to try it?",
          "url": "https://sight4agents.com/how-it-works#early-access",
          "text": "Want to try it?\nsight4agents is in early development. Reach out to hear when early access opens.\nGet early access"
        }
      ]
    },
    {
      "path": "/for-agents",
      "url": "https://sight4agents.com/for-agents",
      "markdown_url": "https://sight4agents.com/for-agents.md",
      "title": "For agents: machine-readable endpoints and MCP for sight4agents.com",
      "description": "How AI agents can navigate, read, and search sight4agents.com through llms.txt, Markdown pages, a JSON content index, search APIs, and a read-only MCP endpoint.",
      "last_updated": "2026-11-12",
      "sections": [
        {
          "id": "overview",
          "heading": "For agents",
          "url": "https://sight4agents.com/for-agents",
          "text": "For agents\nThis site is built the way sight4agents builds sites. Everything a person can read here, an agent can read in a structured form. All endpoints are public, read-only, and need no authentication."
        },
        {
          "id": "endpoints",
          "heading": "Machine-readable endpoints",
          "url": "https://sight4agents.com/for-agents#endpoints",
          "text": "Machine-readable endpoints\n Endpoint Format Purpose \n /llms.txt Markdown Overview of the site with links to every page, following the llms.txt convention \n /llms-full.txt Markdown The full text of every page in a single document \n /<page>.md, e.g. /index.md Markdown Markdown version of any page. You can also send Accept: text/markdown \n /api/content.json JSON Content index: every page with its URL, title, description, and sections \n /search.json?q=mcp JSON Full-text search over all page sections \n /search?q=mcp HTML The same search, rendered for people \n /mcp MCP (JSON-RPC) Read-only Model Context Protocol endpoint over Streamable HTTP \n /.well-known/mcp.json JSON Discovery hints for the MCP endpoint and other machine-readable resources \n /sitemap.xml XML Sitemap of all canonical pages \n /robots.txt Text Crawler policy. AI crawlers are welcome"
        },
        {
          "id": "mcp",
          "heading": "Connect over MCP",
          "url": "https://sight4agents.com/for-agents#mcp",
          "text": "Connect over MCP\nThe MCP endpoint at https://sight4agents.com/mcp speaks JSON-RPC 2.0 over Streamable HTTP (POST, stateless, JSON responses). It provides three read-only tools, and also exposes each page as a resource.\nlist_pages: list every page with its URL, title, and description\nget_page: fetch one page by path (for example /how-it-works) as Markdown or JSON\nsearch: full-text search across the site; returns matching sections with links"
        },
        {
          "id": "guidelines",
          "heading": "Guidelines for agents",
          "url": "https://sight4agents.com/for-agents#guidelines",
          "text": "Guidelines for agents\nPrefer Markdown or JSON over parsing HTML. The content is identical.\nCanonical URLs use https://sight4agents.com with no trailing slash.\nYou are welcome to index, quote, and summarize this site. Please link back to the canonical page.\nEverything here is read-only. In sight4agents products, changes proposed by agents always go to a human for approval."
        }
      ]
    },
    {
      "path": "/about",
      "url": "https://sight4agents.com/about",
      "markdown_url": "https://sight4agents.com/about.md",
      "title": "About sight4agents",
      "description": "Why sight4agents exists and the principles behind it: humans approve changes, open standards like MCP, and content that is readable by people and agents by default.",
      "last_updated": "2026-11-12",
      "sections": [
        {
          "id": "overview",
          "heading": "About sight4agents",
          "url": "https://sight4agents.com/about",
          "text": "About sight4agents\nsight4agents is a project to make websites and applications work well for the AI agents that increasingly read and maintain them, without taking people out of the decisions that matter."
        },
        {
          "id": "principles",
          "heading": "Principles",
          "url": "https://sight4agents.com/about#principles",
          "text": "Principles\nHumans approve: Agents can propose anything, but people decide what goes live.\nOpen standards: Built on open conventions like MCP, Schema.org, sitemaps, and llms.txt rather than proprietary formats.\nReadable by default: One source of content, rendered for people and for machines, so neither gets a lesser version.\nFast and accessible: Lightweight pages that load quickly and work with assistive technology."
        },
        {
          "id": "status",
          "heading": "Status",
          "url": "https://sight4agents.com/about#status",
          "text": "Status\nsight4agents is in early development. This website is itself an example of the approach: it has semantic HTML, Markdown versions of each page, a content index, search, and a read-only MCP endpoint. See For agents. The name, logo, colors, and type are documented on the Brand page."
        },
        {
          "id": "contact",
          "heading": "Get in touch",
          "url": "https://sight4agents.com/about#contact",
          "text": "Get in touch\nQuestions, ideas, or interest in early access? We would like to hear from you.\nEmail sight4agents"
        }
      ]
    },
    {
      "path": "/brand",
      "url": "https://sight4agents.com/brand",
      "markdown_url": "https://sight4agents.com/brand.md",
      "title": "sight4agents brand: logo, colors, type, and voice",
      "description": "The sight4agents brand guide: the site-vs-sight name, the eye-in-a-browser logo, the Ink, Paper, Iris, and Signal color palette, Space Grotesk, Inter, and JetBrains Mono type, and voice guidelines.",
      "last_updated": "2026-11-12",
      "sections": [
        {
          "id": "overview",
          "heading": "Brand",
          "url": "https://sight4agents.com/brand",
          "text": "Brand\nsight4agents is a site with sight, for agents. The brand is drawn from an agent's point of view: a browser window that looks back, a text cursor for a pupil, and a signal color for what has been seen."
        },
        {
          "id": "name",
          "heading": "Name and tagline",
          "url": "https://sight4agents.com/brand#name",
          "text": "Name and tagline\nWrite the name as one lowercase word: sight4agents. Use the numeral 4, never \"for\". In the wordmark the 4 is set in Iris.\nTagline: Sites with sight. Supporting line: \"Websites agents can see, and humans sign off on.\" Short form for small spaces: \"Site, meet sight.\""
        },
        {
          "id": "logo",
          "heading": "Logo",
          "url": "https://sight4agents.com/brand#logo",
          "text": "Logo\nThe mark is a browser window that has become an eye. The almond shape has bracket-like corners (< and >), the iris is Signal lime, and the pupil is a blinking text cursor: the moment an agent reads a page. The third window dot is lit in Signal to show the site is being seen.\n!sight4agents logo: eye-in-a-browser mark with the wordmark, on a light background\n*Primary logo (light backgrounds)*\n!sight4agents logo on a dark background\n*Logo for dark backgrounds*\n!sight4agents mark: an eye inside a violet browser tile\n*Mark (app icon, avatar)*\n!sight4agents outline mark\n*Outline mark (one-color contexts)*\nKeep clear space around the logo equal to the height of the window dots row (about a quarter of the mark).\nDo not recolor the iris, rotate the mark, or set the wordmark in another typeface.\nMinimum size: 20px for the mark, 96px wide for the full logo. Use /favicon.svg below that.\nDownloads: logo.svg, logo-dark.svg, mark.svg, mark-outline.svg, wordmark.svg, apple-touch-icon.png, og.png."
        },
        {
          "id": "colors",
          "heading": "Color",
          "url": "https://sight4agents.com/brand#colors",
          "text": "Color\nFour core colors. Iris carries the brand, Signal marks what has been seen or needs attention, and Ink and Paper do the reading work. Signal is a highlight, not a text color on light backgrounds.\n Color Hex CSS variable Use \n Ink #0E1020 --s4a-ink Text on light, background in dark mode \n Paper #F6F5F0 --s4a-paper Background on light, text in dark mode \n Iris #5B3DF5 --s4a-iris Primary brand color, links, buttons \n Iris Light #A493FF --s4a-iris-light Iris on dark backgrounds \n Signal #C6F432 --s4a-signal Highlights, focus, the iris of the eye \n Slate #575D6E --s4a-slate Secondary text on light \n Mist #E4E2DA --s4a-mist Borders and dividers on light \n Night #161A2E --s4a-night Raised surfaces in dark mode"
        },
        {
          "id": "type",
          "heading": "Typography",
          "url": "https://sight4agents.com/brand#type",
          "text": "Typography\nThree open-source typefaces (SIL Open Font License), all self-hosted as variable fonts.\n Role Typeface CSS variable Sample \n Display: wordmark, headings, buttons Space Grotesk --font-display Sites with sight. \n Body: paragraphs, UI text Inter --font-body Agents propose. Humans approve. Nothing ships without a yes. \n Mono: code, endpoints, the agent view JetBrains Mono --font-mono GET /llms.txt → 200 text/plain"
        },
        {
          "id": "voice",
          "heading": "Voice",
          "url": "https://sight4agents.com/brand#voice",
          "text": "Voice\nPlain and specific. Say what the site does in words a person and a parser both understand.\nTwo audiences, one truth. Never say something to agents that differs from what people see.\nHumans decide. Describe agents as proposing and people as approving.\nLight on puns. The name does the wordplay, so use \"sight\" jokes sparingly."
        }
      ]
    }
  ]
}
