SEO checker

What the checker does

Type a page’s address and press Check. Our server fetches that one page once, with the site’s robots.txt, llms.txt and sitemap. Everything else runs in your browser: 85 checks, three scores and the code to fix what’s wrong. Nothing is saved, and the share link holds only the address. If the page sits behind a login or isn’t live yet, use Paste HTML, which never leaves your device, or Start from scratch to test a title and description before you publish. Your text checks a draft: type or paste it, or open a .docx, .odt, .md, .txt or .html file. It is read on your device and never uploaded.

Checks for each page type

The checker works out what kind of page it is and checks what that kind should show. You can change the type. An article needs a byline, a date, pictures and sources. News also needs a time. A product needs a price, stock, photos, delivery and returns. A recipe needs an ingredients list with amounts, numbered steps, times, servings and a photo. An event needs a date, time, place, a way to register and the price. A local business needs an address, phone and hours. A service page needs what you offer, a price or how you quote, the area you serve, a way to get in touch, proof and answers to common questions. A tool or app page needs the tool on the page, one line on what it does, how to use it, whether it is free and where it runs. A video page needs the video, words around it, a transcript or chapters, a thumbnail and a date. A job posting needs the plain job title, the place (or remote), pay, the job type, how to apply and a closing date. A shop category needs a short introduction, products with prices in the page itself, real links to further pages, breadcrumbs and a clean canonical. An about or profile page needs who you are, a photo, contact details, your experience and links to your profiles elsewhere. Google has no rich result for service or category pages, so their schema is shown as optional. For live pages it also compares the page’s structured data with the required and recommended properties on Google’s structured data pages (updated September 2026), and writes the JSON-LD for that type from your page or draft.

Main content only

Menus, sidebars and footers repeat on every page. By default the word count, top words and reading ease use only the main content: the page’s main or article area, or else its biggest block of text. Switch to Whole page to compare.

Three scores: SEO, AEO and GEO

• SEO is classic search: can Google and Bing read the page, index it and show it well? Titles, descriptions, canonical tags, links, speed and structured data.
• AEO, answer engine optimisation, is about being picked as the direct answer: a featured snippet, “People also ask”, a voice answer or a Google AI Overview. These quote short, clear answers that sit right under a question, and lists and tables.
• GEO, generative engine optimisation, is about AI assistants such as ChatGPT, Claude, Gemini and Perplexity: can their crawlers read the page, can they trust it (a named author, dates, sources), and does it hold facts worth quoting?

Each score runs from 0 to 100: Strong from 85, Good from 65, Needs work from 40. Every check has a weight. A pass earns its points, a warning half, a fail none, and the share is squared, so one slip barely moves a page while several real problems pull it down clearly.

Blockers

Some problems hide a page however good the rest is:

• the page set to noindex;
• the page answering with an error;
• Google or Bing blocked in robots.txt;
• a canonical tag pointing to another website;
• for AI, every AI search crawler blocked, or almost no text without JavaScript.

Any of these caps the score at 25 until it’s fixed, so a hidden page can never look “Good”.

Titles and descriptions

Google says there is no length limit on titles or meta descriptions. It cuts them to fit the screen (Search Central, December 2025 and April 2026). So the checker measures them in pixels, in Arial, the font used for result text: a title fits up to about 580–600 px (roughly 55 to 60 characters), a description about 900–920 px. You can edit both in the Meta tags tab and watch the preview, the checks and the scores change.

AI crawlers and robots.txt

The AI readiness tab reads the site’s real robots.txt for your page and shows each AI crawler as allowed or blocked. The difference matters:

• Search crawlers fetch pages to show and link them in answers. Blocking them hides you there. These are OAI-SearchBot for ChatGPT search, Claude-SearchBot, and PerplexityBot.
• Training crawlers collect text for training AI models. Blocking them keeps your content out of training, a fair choice that doesn’t hide you from answers. These are GPTBot, ClaudeBot, Google-Extended, CCBot and others.

These roles come from the companies’ own documentation (OpenAI and Anthropic, read October 2026).

Structured data and the 2023 change

The checker reads every JSON-LD block, names its type and lists missing properties. In 2023 Google stopped showing HowTo rich results and limited FAQ rich results to well-known government and health sites. That markup does no harm and still describes your questions clearly to other tools, so it counts only as a small plus for answers and AI, never for search. The Make the JSON-LD box writes Organization, Article, FAQ (from your page’s own questions), Breadcrumb and LocalBusiness blocks, filled in from the page.

WhatsApp previews

Most links in India are shared on WhatsApp. Meta’s WhatsApp link preview documentation asks for:

• an absolute og:image address;
• the image under 600 KB and at least 300 px wide;
• the tags within the first 300 KB of the HTML.

Developers report that images under about 300 KB show most reliably. The checker reads your share image’s real size and file weight, and shows the WhatsApp, Facebook and LinkedIn, and X previews side by side.

Selvi’s bakery page

Selvi runs a bakery in Kallakurichi. In October 2026 she checks her cake page and gets SEO 58, AEO 41, GEO 49. The top fixes:

• the title is cut;
• there’s no share image;
• her phone number isn’t a tap-to-call link;
• there’s no LocalBusiness schema;
• her “Do you deliver?” heading has no direct answer under it.

She shortens the title to “Birthday cakes in Kallakurichi, same-day delivery | Selvi Bakes”, adds a 1200 × 630 photo at 180 KB, links the number, pastes the LocalBusiness block from the checker, and answers delivery in two sentences. Her next check reads SEO 91, AEO 82, GEO 84. She copies the fix pack to the person who maintains her site, so nothing is missed.