Link preview, AI visibility, app links, performance and security checker

Check every share before it ships.

Paste a URL to preview its share card on Facebook, WhatsApp, X, Telegram, LinkedIn, iMessage, Discord, Slack, Messenger, Instagram, Snapchat and TikTok, see whether ChatGPT, Claude and Perplexity can read it, check that it opens your iOS or Android app, find what slows it down, and see which security headers are missing. Every report ends with copy-ready fixes. Or ask your AI assistant: Sharelint is also an MCP server for ChatGPT, Claude, Cursor and VS Code.

How it works

Link previews are built by each platform's crawler, which downloads your page's raw HTML and reads its meta tags without running JavaScript. This checker does the same: it fetches your page with the real crawler User-Agent, follows redirects, parses og: and twitter: tags, downloads the preview image to measure its dimensions and weight, and applies each platform's own rules for layout, fallbacks and limits.

The AI visibility check reads your robots.txt rules for each AI crawler, requests the page with each bot's User-Agent to catch blocks at your server or CDN, and shows the text AI tools get from your raw HTML.

The app links check reads your apple-app-site-association and assetlinks.json files, tests whether the exact URL opens your app, and compares them with the copies Apple's CDN and Google's Digital Asset Links service have cached, since those copies are what phones use.

The performance check times the server response as a regular browser, reads the HTML for render-blocking scripts and stylesheets, the main image and images without dimensions, and shows real-user Core Web Vitals from the Chrome UX Report when Google has data for the site.

The security check reads what every browser receives: HTTPS and the http:// redirect, Strict-Transport-Security, Content-Security-Policy, clickjacking protection, cookie flags, http:// files on https pages and security.txt. It is passive, so it never attacks or scans the site.

Supported platforms

Use Sharelint from your AI assistant

Sharelint is a remote MCP server, so ChatGPT, Claude, Cursor, VS Code and other MCP clients can run its checks for you. Ask "why does my link show no image on WhatsApp?" or "can ChatGPT read my site?" and the assistant checks the live page and walks you through the fixes. No key or sign-up needed.

https://sharelint.com/api/mcp

Tools: check_link_preview, check_ai_visibility, check_app_links, check_performance, check_security and check_url for a full audit.

MCP server setup guide, with each tool and example questions.

API for CI, scripts and AI agents

Every check is available as plain HTTP, with no key needed. Ask for Markdown to get a report you can paste into a pull request or hand to an LLM.

curl "https://sharelint.com/api/preview?url=https://example.com&format=md"

See openapi.json for the schema and llms.txt for an LLM-ready summary of this site.

Frequently asked questions

Why does my link show no preview, or the wrong image?

Platforms build previews from the og:title, og:description and og:image tags in the raw HTML of your page. Their crawlers do not run JavaScript, so tags added by client-side code are invisible to them. The image must be an absolute URL, publicly reachable, and within each platform’s size and format limits.

Why does WhatsApp not show my og:image?

WhatsApp skips images that are too heavy (keep them under 300 KB, and never above about 600 KB), served over plain http, or in formats such as SVG or AVIF. Use a 1200×630 JPEG or PNG and an absolute https URL.

How do I refresh a cached preview?

Facebook, Messenger and Instagram: use the Facebook Sharing Debugger and click Scrape Again. LinkedIn: use Post Inspector. Telegram: send the link to @WebpageBot. X, WhatsApp, iMessage, Discord and Slack have no public refresh tool, so add a query string such as ?v=2 to make them treat it as a new URL.

What image size works on every platform?

A 1200×630 pixel JPEG or PNG under 300 KB, served from an absolute https URL, with og:image:width and og:image:height set. Keep important content in the centre, because square and small layouts crop the sides.

Can ChatGPT, Claude and Perplexity read my page?

The AI visibility check tells you. It reads your robots.txt rules for each AI crawler, requests the page with each bot’s User-Agent to catch CDN blocks, and shows the text they get from your raw HTML. AI crawlers do not run JavaScript, so a page rendered in the browser can look empty to them. Blocking training crawlers such as GPTBot does not remove you from AI answers; blocking search crawlers such as OAI-SearchBot does.

Can I use Sharelint from ChatGPT, Claude or Cursor?

Yes. Sharelint is a remote MCP server at https://sharelint.com/api/mcp, with no key needed. Add it as a custom connector in ChatGPT or Claude, or as an MCP server in Cursor, VS Code or Claude Code. Then ask your assistant to check a link’s preview, AI visibility, app links, performance or security, and it runs the check on the live page and explains the fixes.

Why does my link open the browser instead of my app?

iOS opens a link in your app only when apple-app-site-association on your domain lists the path, and Android only when assetlinks.json verifies your app. Both files must return HTTP 200 as JSON with no redirects. iOS reads the copy on Apple’s CDN, which can lag a day behind your file, and Android verifies through Google’s Digital Asset Links service. The app links check tests all four, plus the exact URL you enter.

Does Sharelint scan my site for vulnerabilities?

No. The security check is passive: it reads what every browser receives, such as HTTPS, Strict-Transport-Security, Content-Security-Policy, X-Frame-Options, cookie flags, http:// files on https pages and security.txt, and gives you the headers to add. It sends no attack payloads, so it cannot find bugs in your code or outdated plugins. Use a penetration test for that.

Can I check localhost or a staging site behind a login?

Use the Sharelint browser extension for Chrome (see /browser-extension/). It runs every check from your own browser, so localhost, staging and VPN-only pages work. Like the real crawlers it sends no cookies, so pages behind a login are checked as a logged-out visitor sees them. The website itself fetches only public pages.

Is there an API for scripts, CI and AI agents?

Yes. GET /api/preview?url=<url> returns JSON. Add format=md to get a Markdown report that is easy to paste into a pull request or hand to an LLM. See /llms.txt and /openapi.json.