Introducing SEO Intelligence — audit any site in under a minute

Search doesn't look the way it used to. When your prospects want to evaluate a product, shortlist a vendor, or answer a nagging question, they aren't always typing a query and clicking blue links. They're asking an AI — and reading the synthesised answer. If your brand isn't inside that answer, it doesn't exist for them. You can rank number one on Google and still lose.

This has created a split-brain problem for marketing teams. Half your pipeline is still driven by classical SEO — titles, links, Core Web Vitals, the familiar checklist. The other half is driven by something genuinely new: whether large language models have enough of the right signals to pick you as the source they quote. The old tools are great at measuring the first half and blind to the second.

MarketOS's new SEO Intelligence module is our answer. Point it at any URL — yours, a competitor's, a site you're auditing for a client — and in under a minute you get a prioritized report that covers both halves of the picture. The deep version also crawls the rest of the site, measures real Core Web Vitals from Google's own data, fetches HTTP headers, and actually tests whether AI answer engines cite the domain when real prospects ask buyer-intent questions.

How it works, in a sentence

You paste a URL. A 4-step wizard asks what you want to rank for, who your competitors are, and what the goal is. You press "Run audit." A panel shows the scan running live — fetching the page, checking robots.txt, discovering your sitemap, measuring performance, crawling pages, asking AI engines live questions. When it's done, a score ring reveals your overall grade, five sub-scores stagger in with a short narrative from the AI strategist, and then you land in the action center: a tabbed workspace full of fixes you can apply today.

That's it. No CSV exports, no waiting for overnight queues, no 40-tab dashboards.

The five pillars we grade on

Every audit produces an overall score (0-100) and five sub-scores. Each sub-score exists because the work to improve it is genuinely different. Mixing them together would hide the story.

The overall score is weighted with GEO and On-page carrying the most weight — because in 2026, that's where the ranking and citation battles are won or lost.

The GEO pillar — why this is the one most tools miss

"GEO" stands for Generative Engine Optimization. It's the craft of making sure LLM-powered search — ChatGPT, Perplexity, Google AI Overviews, Claude, Copilot — picks your site when someone asks a question your brand should be answering.

Here's the uncomfortable part: ranking on Google and being cited by AI are only weakly correlated. A brand can hold the top spot on page one and still be completely absent from ChatGPT's answer to the same question. AI engines weight different things. They care about schema, entity density, concrete statistics, clear authorship, and whether your content is structured for quick extraction (a TL;DR lead, question-form headings, FAQ patterns). They care less about backlinks and keyword density.

Our audit measures these specific things — and then goes further.

Live AI citability tests — across four engines

During a deep audit, MarketOS generates 8 realistic buyer-intent queries based on your brand and target keywords. Then it sends each query to every AI engine you've configured, with real-time web search enabled, and parses the sources each engine cites. Four engines are supported today:

• Google AI Overviews — via Gemini with Google Search grounding. Same retrieval pipeline Google uses for AI Overviews.
• ChatGPT — via OpenAI's Responses API with the web_search tool. Uses Bing under the hood.
• Perplexity — via the Sonar API, which returns structured citations for every response.
• Claude — via Anthropic's Messages API with the web_search tool. Uses Brave Search.

Each engine has a different retrieval pipeline, which means the same page can be cited by one and ignored by another. That's why testing all four matters — one-engine coverage is still one-engine blind.

For each query across each engine, you see concretely:

• Your aggregate AI visibility rate (average across engines), plus a per-engine breakdown so you can tell if you're strong on AI Overviews but invisible on Perplexity.
• Which specific queries cited you, which didn't, and the AI's actual answer for each one.
• Which competitor domains won the queries you lost — so you can study what they're doing right.
• The most-dominant domains across all queries, giving you a map of the "AI content landscape" in your category.

During the scan itself, a live per-engine progress block ticks up in real time — you can watch Gemini finish eight queries in nine seconds while ChatGPT's web_search tool works through its pipeline. No black box.

You're not guessing whether your GEO work is paying off. You're watching it happen, engine by engine.

Designed for the way you actually work

We built the interface the way we'd want to use it. One question per panel. Progress dots at the top so you know where you are. Everything keyboard-driven — press Enter to advance, Esc to go back. When the scan runs, the ticker shows what's happening in real time: Fetching response headers… Measuring Core Web Vitals… Crawling 7 of 15 pages…

When the audit finishes, you don't get dumped into a wall of tables. You get a cinematic reveal — your score counts up, the five sub-rings stagger in, and the AI strategist writes you a 4-sentence summary that names your single biggest leverage point. Then you enter the action center, where everything is organized by tab:

• Issues — a prioritized list by severity and impact. Filter by category. Each card expands to reveal the exact fix.
• Page — eleven groups of per-page signals, from OpenGraph to E-E-A-T heuristics to structured-data samples.
• Site — what the crawler found across your sitemap: thin pages, duplicate titles, pages missing H1s, schema coverage rates.
• Performance — mobile and desktop Core Web Vitals side by side, flagged against 2026 thresholds.
• AI Citability — every test query, the AI's answer, the sources it cited, and whether you were one of them.
• Headers — security headers, compression, CDN and platform detection, TTFB, and the full raw response header table.
• Artifacts — copy-paste-ready fixes: a drop-in <head> block, Organization JSON-LD, a starter llms.txt.
• Monitor — (coming soon) weekly re-audits with email reports through your existing Resend integration.

Quick audit vs deep audit

Every audit does the same core work: fetch the page, parse the HTML into 85+ signals, check robots.txt, sitemap.xml, llms.txt, and security.txt, score the page, build a prioritized issue list, and write an executive summary.

A quick audit stops there. It takes roughly 8 seconds. Perfect for a rapid sanity check, a client call, or a first pass on a big list of competitor domains.

A deep audit (the default, and recommended) adds four things that genuinely change the diagnosis:

• A multi-page crawl of up to 15 same-site URLs, discovered from your sitemap or the homepage's links. Detects duplicate titles, thin pages, missing H1s across templates, schema coverage rates — systemic issues you can't see from one URL.
• Real Core Web Vitals from Google's PageSpeed Insights API, using the Chrome UX Report's actual field data when available. Mobile and desktop, graded against the 2026 "good" thresholds.
• Response headers fetched via a secure MarketOS edge function — security headers, CDN detection, platform detection (Next.js, WordPress, Shopify, Webflow, etc.), and compression confirmation.
• Live AI citability tests, described above.

Deep audit takes about 45 seconds. It delivers roughly 10× more diagnosis. Run it the first time you look at any site.

It also upgrades your blog generator

This was the second half of the project. MarketOS has always had a blog generator with 100+ content types across 11 categories — standard blog posts, FAQs, comparison articles, listicles, AEO answer pages, and so on. Now there's a toggle in SEO Mode called Include SEO pack, and when it's on, every generated post comes with everything you actually need to ship:

• A properly-sized title tag (40-60 characters).
• A meta description in the 140-160 character sweet spot for SERP CTR.
• A clean URL slug.
• A TL;DR block that AI answer engines can lift verbatim — the single biggest citability lever.
• 3-5 FAQ pairs extracted from the post, ready to render as a visible FAQ block and as FAQPage schema.
• An OG image prompt (use it directly in MarketOS image generation).
• Internal link suggestions as anchor phrases you can place throughout the post.
• The identified primary and secondary keywords so you can verify the post is optimized for what you wanted to rank for.
• Article JSON-LD — drop it into a <script type="application/ld+json"> tag on the published page.
• FAQPage JSON-LD when the post includes FAQs — unlocks FAQ rich results in Google SERPs.

All of it appears in a collapsible "SEO pack" panel above the rendered post. Every field has a copy button. Nothing to retype, nothing to forget.

Who this is for

It's for you if any of these describe your week:

• You're responsible for a site's organic visibility and you're losing patience with legacy tools that don't even measure whether ChatGPT knows your brand exists.
• You manage SEO for multiple clients or multiple brands, and you want to run a full audit on a new engagement in 45 seconds instead of a day.
• You're a founder who ships a lot of content and wants to know whether the content is actually doing its job — not just whether it was published.
• You're evaluating a competitor and want hard data: how does their site score, what's in their schema, where are they getting cited?
• You're about to rewrite a page and want an honest diagnosis of what it currently does poorly, not a vibes-based critique.

Pre-audit cost transparency

Because the audit uses your own provider keys for everything except Gemini, we show you what an audit is about to cost before you run it. On the goal step of the wizard, a live estimate tile lists every engine you've configured with a rough per-audit spend:

• Gemini — "Included in your plan" if you're a subscriber, or ~$0.02 on your own key.
• OpenAI (ChatGPT) — ~$0.10 per audit. 8 web_search calls on GPT-5.4-mini.
• Perplexity — ~$0.05 per audit. 8 Sonar calls.
• Claude — ~$0.08 per audit. 8 web_search_20250305 calls on Claude Haiku 4.5.

Nothing is hidden. You always know what each run will put on your provider bills before you commit to running it, and you can toggle engines off if you only want a cheap subset.

How the plans work

MarketOS comes in three flavors. Pick whichever matches how you want to engage:

• Monthly subscription — $19.99/mo. Run up to 15 deep audits per month on the MarketOS Gemini key (we handle that cost). Other engines are BYOK. Great for trying the product without upfront cost.
• Lifetime purchase — $199 one-time. Includes the Windows + macOS desktop apps and unlimited audits on your own Gemini key. Lifetime is strictly BYOK — every engine uses your own API keys, your data never leaves your machine, and you pay nothing recurring. Best long-term value.
• Trial / exploration — free. Sign in and walk through the full wizard without running an audit. See the interface, the scoring pillars, the content-category tree. Useful for evaluating whether the product fits your workflow.

No matter which route you pick, the behavior is identical once you're running audits. Same features, same scoring, same UI — the differences are all about who pays for the Gemini calls.

Security — paid attention

We care a lot about this. Quick tour of what's in place:

• SSRF-hardened server fetches. Any URL the audit fetches server-side is validated against private IP ranges (RFC1918, loopback, link-local, CGNAT), non-HTTPS schemes, embedded credentials, and known dangerous ports. We also do a DNS-over-HTTPS lookup before fetch to catch DNS rebinding. Your "audit this URL" can't become our "hit an internal service."
• OS-keychain encrypted keys on desktop. The Windows and macOS apps now wrap your saved API keys with Electron's safeStorage, which uses the macOS Keychain, Windows DPAPI, or Linux libsecret. Keys are encrypted at rest — a filesystem snapshot of your user data directory doesn't leak them. First launch after this update silently migrates keys out of plaintext localStorage into the encrypted store.
• BYOK by default. OpenAI, Perplexity, and Anthropic keys are sent directly from your browser/app to each provider. MarketOS servers never see them, never log them, never proxy them. Settings tooltips call out that browser-hosted keys are visible in the network tab so you can decide on scope appropriately.
• Origin-locked CORS on MarketOS APIs. Our backend endpoints now reflect only allowlisted origins, so a malicious site can't CSRF a signed-in subscriber's bearer token.
• Redacted errors. Upstream provider errors are sanitized before we echo them — anything matching a known API-key shape or Bearer token is stripped before it reaches your screen or a log.
• Monthly spend ceiling for subscribers. The $19.99/mo tier is hard-capped at 15 audits per month server-side; we can't accidentally bill you for more. When you hit the cap you're told immediately, with the reset date, and the next audit's options: add your own Gemini key for unlimited use or wait until the 1st of next month UTC.
• Responsible disclosure. /.well-known/security.txt is live with support@driftrail.com as the contact. Found something? Tell us — we respond within 72 hours.

For the paranoid: we also built an internal provider-drift canary (npm run canary:providers) that we run before every release to confirm each provider's API shape hasn't changed under us. If ChatGPT's web_search tool breaks or Anthropic renames web_search_20250305, we catch it before you do.

The audit → content loop

The audit's job is to tell you what's broken and what's missing. The Content Studio is where you fix it. We built them to pair.

Run an audit, find that your FAQ coverage is thin and you lack an llms.txt. Hop over to Content Studio with one click, generate an FAQ page tuned to your brand voice — and because the Studio's SEO pack is built-in, every generated post comes with:

• Title tag (40-60 chars, keyword-forward)
• Meta description (140-160 chars, CTR-tuned)
• URL slug
• TL;DR block (the AI-answer-engine citability lever)
• 3-5 FAQ pairs extracted from the post
• OG image prompt you can feed straight into MarketOS image gen
• Internal-link suggestions as anchor phrases
• Article JSON-LD schema, ready to paste into your CMS
• FAQPage JSON-LD schema when the post includes FAQs

The audit surfaces a gap → you generate content to fill it → the content ships with the schema the audit told you was missing. One tool, two halves, closed loop.

What's next

A few features are coming that we scoped out of the initial releases to ship the core quickly:

• Scheduled weekly re-audits with email reports comparing this week's score to last week's — delta view, new issues, resolved issues. Uses our existing Resend + Netlify scheduled-functions integration.
• Competitor comparison mode that runs the audit on your site and up to 3 competitors in parallel and produces a diff view highlighting the gaps.
• Pillar & cluster builder — pick a pillar topic, MarketOS plans 8-12 interconnected cluster posts with the internal links pre-wired.
• "Rewrite my page" — paste a URL, MarketOS fetches it, analyzes it against the top SERP results, and produces an improved version with the gaps filled.
• Google Search Console integration for authoritative impressions/clicks/CTR data alongside the audit.
• Public shareable audit links so agencies can send read-only reports to clients and prospects.

If one of these is the one you've been waiting for, let us know. Roadmap is community-weighted.

How to use it today

If you already have MarketOS, update to the latest version (v2.2.0 or higher). You'll see a prominent SEO Intelligence card in the sidebar — the product's flagship module, marked with a small "Flagship" pill. Click it, paste a URL, answer four quick questions on a stepped wizard, and watch it run.

Three ways to run audits:

• Just sign in to try — click the SEO Intelligence button, walk the wizard; on Run Audit you'll see a polished upgrade screen with three options for connecting Gemini.
• Grab a $19.99/mo subscription for 15 audits/month on MarketOS's Gemini key, plus the whole marketing suite.
• Own it forever for $199 one-time with the desktop app, unlimited audits on your own keys, and privacy-first local-first operation.

You also get the rest of the MarketOS suite alongside SEO Intelligence: brand-aware image generation with 20+ models, caption and hashtag generation tuned to your brand voice, long-form content in 100+ formats across 11 categories, and a full agentic REST API for programmatic use. Everything pairs.

Go audit something.

We built SEO Intelligence because we were tired of running three tools to diagnose what one tool should tell us, and because the one tool that would tell us everything didn't exist. Now it does — and it ships with everything else you need to do modern marketing in the same workspace.

Go audit something.