> ## Documentation Index > Fetch the complete documentation index at: https://docs.yournarrative.ai/llms.txt > Use this file to discover all available pages before exploring further. # Technical SEO in Narrative > Why technical health decides whether search engines and AI answer engines can use your pages at all, and how Narrative measures it Every engine that could cite your brand starts the same way: a crawler fetches a page, reads it, and decides whether it can use what it found. Google does this. So do the crawlers behind ChatGPT, Perplexity, Gemini, Claude and Copilot. The content on the page only matters once that step succeeds. Technical SEO is the discipline of making that step succeed. Narrative treats it as part of answer engine optimization rather than a separate product, because the same faults that hide a page from Google hide it from every answer engine, and a few faults hide it from answer engines alone. ## Why it matters more for answer engines Search engines have spent two decades learning to tolerate messy sites. Googlebot renders JavaScript, follows long redirect chains, and guesses at canonical URLs. Answer engines mostly do not: * **Most AI crawlers do not run JavaScript.** A page whose content arrives only after a script runs is empty to OAI-SearchBot, PerplexityBot and Claude-User. Nothing about the copy can compensate. * **robots.txt is a hard wall.** A brand that disallows `GPTBot` or `PerplexityBot` is absent from those engines, however good its pages are. Many sites did this in 2023 and never revisited it. * **Answer engines quote passages, not pages.** They need clean, extractable text, one clear answer per section, and metadata that says what the page is. Noise in the markup becomes noise in the answer, or no answer. * **Fewer crawls, less patience.** An AI crawler visits a page less often than Googlebot does. A redirect chain, a slow response or a buried page costs more when there are fewer chances to get it right. Narrative's audit library is built around this. It carries the technical checks Google documents, plus the ones the answer engines need and Google does not check for you. The gates in the library, the checks that can make a brand invisible outright, are the answer-engine ones: server-side rendering without JavaScript, and AI crawler access in robots.txt. ## What Narrative does **1. Crawl your whole site** Narrative's own crawler, [NarrativeBot](https://docs.yournarrative.ai/technical-seo/narrativebot.md), walks your site the way an engine does: it reads robots.txt and your sitemaps, follows internal links, honours crawl delays, and records every page, link, redirect and response. It does not sample. From the result it builds your internal link graph. **2. Audit every page against the library** Each page is judged against the [audit library](https://docs.yournarrative.ai/technical-seo/audit-library.md): per-page checks that read the page itself, and whole-site checks that read the graph, because a page cannot know what links to it or that another page is a copy of it. Field performance comes from [Chrome's real-user data](https://docs.yournarrative.ai/technical-seo/core-web-vitals.md). **3. Show you the state of every page** **Site Health** (under Context in the sidebar) lists every page with what is wrong with it, what it served, how deep it sits, how much internal authority reaches it, and how fast it was. Group by page, by issue, or walk the link graph. Site-wide vitals sit on the Analyze page. **4. Turn findings into Repairs and Fixes** From Site Health you open a [Repair](https://docs.yournarrative.ai/technical-seo/repairs-and-fixes.md): one family of issues, every page affected, plain-language steps. Each page inside it is a Fix that tracks its own state. Repairs live under **Projects** next to your campaigns. **5. Re-crawl and close what is fixed** Run the crawl again whenever you want. A Fix the audit can no longer see closes itself. One that still shows is recorded as still observed, with the date. You never have to tick a box to say a problem is gone. ## How the pieces fit | Where | What you see | | - | - | | **Context → Site Health** | Every page, its findings, link-graph position, response and speed. Start a crawl. Open a repair. | | **Analyze → Site vitals** | Core Web Vitals for the site, from Chrome's field data. | | **Work → Projects** | Repairs beside campaigns, with open and done tabs. | | **Repair detail** | The problem in plain words, why it matters, the steps, and every page as a Fix. | ## Severity Every check in the library carries a default severity, which is how Narrative orders what to fix first: | Severity | Meaning | | - | - | | **Gate** | The brand is invisible to at least one engine until this is fixed. Nothing else on the page can compensate. Two checks are gates: rendering without JavaScript, and AI crawler access. | | **High** | The page is reachable but materially weakened: unindexable, orphaned, duplicated, or stripped of the schema that identifies the brand. | | **Medium** | The page works but leaves signal on the table: shared titles, buried pages, slow paint, redirect chains. | | **Low** | Hygiene. Worth doing once the rest is clean. | ## Scope and honesty Narrative only reports what it measured. A check that is seeded in the library but not yet wired to a checker records nothing rather than a pass, and the library page marks those as **Coming**. A page whose body could not be fetched is reported as such, not graded. Core Web Vitals come from real Chrome users, so a page too small to have field data simply has none, and that is said rather than inferred. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.