Documentation

SEO Captain works on any http:// or https:// page, including pages behind a login (since the content script reads the already-rendered DOM in your browser). It does not work on browser-internal pages (chrome://), the Chrome Web Store, PDF viewer pages, or pages that explicitly block all content script injection.

No — all analysis runs locally in your browser. Page data is never sent to echozat.com or any analytics service. The only external call the extension makes is to api.anthropic.com when you explicitly click “Generate with Claude AI” — and even then, only if you’ve configured your own API key. You are in full control.

Yes. The extension itself is free. The only cost involved is if you use AI generation — that requires an Anthropic API key billed by Anthropic directly (pay-per-use). The JS auto-generation engine (titles and descriptions) is completely free and requires no API key.

This usually happens for one of these reasons: (1) the page is a browser-internal page like chrome://settings or the Chrome Web Store — these block all extensions, (2) the page uses a Content Security Policy that prevents script injection, or (3) the extension was just installed/updated and the tab hasn’t been refreshed yet. Try hard-refreshing the tab (Cmd+Shift+R / Ctrl+Shift+R) and reopening the extension.

The content script loads once when a tab opens and is guarded against re-injection. To get fresh data after making changes to a page, hard-refresh the tab first (Cmd+Shift+R / Ctrl+Shift+R), then open SEO Captain. Simply closing and reopening the popup is not enough — the tab itself needs to be refreshed.

Google measures title display width in pixels, not characters. Wide letters like W, M, and uppercase text consume more space than narrow letters like i, l, or lowercase text. A 58-character title with many wide characters can exceed Google’s ~580px display budget. If you’re seeing truncation, try shortening to 50–54 characters to create a safe margin, especially if your title has many capital letters or wide punctuation.

No tool can prevent Google from rewriting titles and descriptions — that’s entirely Google’s decision. However, Google is less likely to rewrite them if they are: descriptively accurate (match the page content), within the character limits, and not keyword-stuffed. SEO Captain helps you meet all those criteria, which significantly reduces the chances of rewriting. But it cannot guarantee Google will use your text as-is.

Not at all. The <meta name="keywords"> tag has been officially ignored by Google since 2009, and by most major search engines since around the same time. Leaving it blank is perfectly fine. The only edge cases where you might still want it are: certain minor search engines, some CMS systems that generate sitemaps or internal search indexes from it, or legacy technical requirements on specific platforms.

It depends. A canonical tag is most valuable when: (1) the same content is accessible via multiple URLs (e.g., with and without trailing slash, with/without URL parameters, HTTP vs HTTPS), (2) you have very similar content on multiple pages, or (3) your site accepts tracking parameters that create URL variations. For a simple page with a unique URL, no canonical is still perfectly fine. Google will often infer the canonical itself.

SEO Captain specifically reads <script type="application/ld+json"> blocks — the JSON-LD format. It does not detect schema implemented as Microdata (inline HTML attributes like itemprop) or RDFa. If your structured data uses one of those older formats, the indicator will show “No schema” even though schema is technically present. Google supports all three formats, but JSON-LD is now the recommended approach.

Focus on the highest-weight failing or warning checks first. Sort by points: HTTPS (8), H1 (8), meta description exists (8), image alt text (8), title exists (6), title length (6), description length (6), JSON-LD (6), OG tags (6), viewport (7). A page with all these passing easily reaches 85+. Common quick wins at the “Good” level are: adding missing OG tags, fixing title/description length, and ensuring all images have alt attributes.

The 300-word minimum is a general guideline, not an absolute rule. Short landing pages, contact pages, and homepages that convert well at low word counts are perfectly fine. The check is designed to catch unintentionally thin pages where content just hasn’t been written yet. You can safely ignore this check for pages where brevity is a deliberate design choice.

If you have a <meta name="robots" content="noindex"> or nofollow directive set (possibly added by your CMS automatically), the audit will warn about it even if you don’t intend to block indexing. Check your page source to confirm whether the tag is actually present and what value it’s set to. Many CMS platforms add noindex to draft, staging, or search results pages by default.

No — and any tool that claims otherwise is misleading you. The SEO Captain score measures on-page technical and structural signals. Rankings also depend on factors outside the page: domain authority, backlink profile, Core Web Vitals, user engagement signals, topical relevance, competition, and Google’s constantly-evolving algorithms. A high score means your page’s on-page fundamentals are solid — which is a necessary but not sufficient condition for ranking well.

Not necessarily. First, check if the low score is appropriate for your audience. Technical documentation, medical content, legal text, and academic writing is expected to be complex — a Flesch score of 30 is fine for a whitepaper aimed at PhDs. But if you’re writing for a general consumer audience and your score is low, the most effective improvements are: breaking long sentences into shorter ones, replacing jargon with plain language, and converting complex compound sentences into two simpler ones.

The table shows the top 10 keywords by frequency after filtering ~80 common stop words. Your keyword might not appear because: (1) it’s less frequent than the 10 keywords shown, (2) it’s in the stop word list (very common words), (3) it’s a multi-word phrase (the tool analyzes individual words, not phrases), or (4) the body text used for analysis is capped at 2,000 characters — if the page is very long, some content may not be captured.

SEO Captain reads publication dates from machine-readable structured signals: <meta property="article:published_time">, itemprop="datePublished", <time pubdate>, and similar patterns. A date visible in the page’s body text or displayed by a custom component that doesn’t use these standardized attributes won’t be detected. To make dates machine-readable (and improve Google’s understanding of freshness), add the appropriate schema.org or Open Graph date meta tags.

File sizes come from the browser’s Performance Resource Timing API. Three situations cause a blank: (1) cross-origin images without a Timing-Allow-Origin response header — browsers block their timing data for privacy, (2) images served from cache where transferSize = 0 and no encodedBodySize is available, (3) inline data URIs — these have no network request to measure. Blank size is not an error; it just means timing data isn’t available for that particular image in this browser session.

Several reasons can cause a mismatch: (1) Images loaded by JavaScript after the initial DOM render may not be captured if the extension opened before they loaded — hard-refresh and reopen. (2) CSS background images in external stylesheets are not captured (only inline style attributes are); images set via CSS classes in linked stylesheets won’t appear. (3) Images inside iframes are in a different document context and cannot be accessed.

A few possibilities: (1) The server returns a 404 for HEAD requests but works for GET requests — this is a server misconfiguration but not uncommon. (2) The page requires JavaScript to render and the HEAD request hits a pre-render that isn’t ready. (3) Rate limiting — if you checked many links quickly, the server may have temporarily blocked the requests. (4) Authentication — the link requires a session cookie that the HEAD request doesn’t carry. Try using the “Re-check” button and visiting the URL directly to compare.

The 25-link cap prevents the extension from overwhelming the server with simultaneous HEAD requests, which could trigger rate limiting or look like a denial-of-service attempt. All 25 requests run in parallel for speed. For larger link audits on production sites, a dedicated crawl tool like Screaming Frog, Sitebulb, or Ahrefs Site Audit is more appropriate.

Sign up at console.anthropic.com, create a new project, and generate an API key from the Keys section. New accounts receive a free credits tier. Beyond that, Anthropic charges per token (input + output) on a pay-per-use basis. Each SEO Captain AI generation call uses a small amount of tokens — typically well under 1,000 input tokens and 400 output tokens per call.

Your API key is stored in chrome.storage.local, which is scoped exclusively to the SEO Captain extension. Web pages cannot access it; other extensions cannot access it. It is only transmitted to api.anthropic.com when you click “Generate with Claude AI”, over HTTPS. It is never sent to echozat.com, never logged, and never stored anywhere except your own browser. As a best practice, you can set usage limits on your Anthropic API key from the console to cap any potential misuse.

Common causes: (1) Invalid API key — double-check the key in Settings; it should start with sk-ant-api03-. (2) Insufficient credits — log into console.anthropic.com to check your balance. (3) Rate limiting — Anthropic’s API has per-minute request limits; wait a few seconds and try again. (4) Network issue — check your internet connection. If the problem persists, the Anthropic status page at status.anthropic.com will show any ongoing incidents.

The AI generates suggestions based on what it receives: URL, title, description, H1, top H2s, and a 1,200-character body text slice. Pages where the most relevant content appears deeper in the body (beyond the first ~1,200 characters) may yield less accurate suggestions. Also, pages with very little readable body text (mostly images, videos, or JavaScript-rendered content with no server-side rendered text) give the AI less to work with. For those pages, the JS auto-generation engine using the heading structure often produces more usable results.

Paste the copied JSON-LD (which includes the <script type="application/ld+json"> wrapper) inside the <head> section of your HTML page. In a CMS like WordPress, most SEO plugins (Yoast, Rank Math) have a field for custom schema; paste it there. For Shopify, use the Additional scripts section in the theme editor or a custom app. After adding it, validate it with Google’s Rich Results Test to confirm it’s correctly structured.

Snapshots are stored in chrome.storage.local, which is tied to your Chrome profile and the extension. They can be lost if: (1) you used the “Clear all snapshots” button in Settings, (2) you uninstalled and reinstalled the extension (local storage is deleted on uninstall), (3) you cleared Chrome’s site data with an overly broad setting that included extension storage. Snapshots are not synced across devices — they are local to the browser where they were saved.

Not currently — snapshots can be viewed within the extension and deleted individually, but bulk export to CSV/JSON is not available in version 1.0. The HTML Export Report (Export Report button in Tools) exports the current page’s full analysis as a self-contained HTML file you can share, print, or archive.

No. There is no analytics, telemetry, or usage tracking of any kind in the extension. It does not know what pages you’ve opened it on. No data is ever sent to echozat.com. The only external network call the extension can make is to api.anthropic.com, and only when you explicitly click “Generate with Claude AI”.

activeTab — to inject the content script into the current tab and read its DOM. storage — to save your API key and snapshots in local storage. tabs — to open links (Robots.txt, Sitemap, broken link results, GSC) in new tabs. Host permission for api.anthropic.com — to make direct API calls from the browser popup for AI generation. No other permissions are requested.

The content script is injected into pages automatically (because it’s declared as a content script in the manifest for immediate availability), but it only reads and sends data when the popup sends an explicit analyzePage message — which only happens when you open the extension popup on a page. There is no background monitoring, passive data collection, or reading of pages you haven’t explicitly opened the extension on.