An independent guide · Simple visitor counters for every kind of website
The humble visitor counter is still a practical tool. For a personal homepage, a hobby blog or a forum profile, a small number that says “people have been here” is often all you need – no analytics suite required.
The catch is that many hosting platforms restrict JavaScript, and some places, like forum signatures, only accept an external image. This guide explains how JavaScript counters and HTML-only image counters work, which one fits where, and what each one can honestly measure.
The second button leads to counter.gratis, an affiliated service. See the disclosure.
Run in the visitor’s browser. More context, but they only work where scripts are allowed.
A plain <img> tag. Works almost anywhere images do, but counts server requests, not true pageviews.
Caching, blockers, proxies and bots mean every counter is an estimate. Treat the number as a trend.
01
Both approaches end with a request to a counter server. The difference is who builds that request: a script running in the browser, or the browser’s ordinary image loader.
| Aspect | JS JavaScript-based counter | IMG HTML-only image counter |
|---|---|---|
| Installation requirements | Paste a <script> snippet into the page or template. The platform must allow custom JavaScript, which many free website-builder plans don’t. |
Paste an <img> tag, Markdown image or BBCode [img]. The platform only needs to allow external images. |
| Pageview counting | The script runs on each page load and can add a cache-busting value to its request, so repeat views are usually recorded. It can also count navigation in single-page apps. | Counts each image request that reaches the counter server. Cached or proxied copies aren’t counted, and duplicate loads may be counted twice. |
| Referrer information | Can read the current page address and document.referrer, the page the visitor came from, subject to the referring site’s policy. |
Only receives the Referer header of the image request. Modern browsers usually send just the origin cross-site, e.g. https://example.com/ without the path, and proxies often send none. |
| Browser and screen information | Can read screen size, viewport, language and time zone, if the service chooses to collect them. | Limited to request headers: user-agent string, accepted languages and IP address. There’s no way to learn screen size. |
| Compatibility | Self-hosted sites, static hosts and platforms that permit scripts. Blocked on many free builder plans, forums and profile pages. | Very broad: forums, profiles, restrictive builders and Markdown pages. Fails where external images are stripped, rehosted or blocked. |
| Image caching | Mostly not an issue: the script controls its request and can make each one unique. | The main weakness. The counter server can send no-cache headers, but browsers, CDNs and image proxies don’t always respect them. |
| Limitations | Doesn’t count visitors with JavaScript disabled or blocked by content blockers, or pages whose Content Security Policy forbids third-party scripts. | Doesn’t count visitors whose images are blocked or served from a cache. It can’t tell people from bots well, and estimating unique visitors is harder. |
02
What matters is not whether a platform is “compatible” but what it lets you embed: a script, raw HTML, an image from an external URL, or nothing at all. Plans and policies change, so check the platform’s current documentation and test on a published page.
Embed with the Image block “Insert from URL”, or an <img> tag in a Custom HTML block.
<img> tag in a Code Block or a Markdown block may work. Make sure the image is still loaded from the counter server and hasn’t been uploaded as a copy.Don’t download and re-upload the counter image. A copy on Squarespace’s servers will never change.
Official documentation: Squarespace: custom code documentation
<img> inside an Embed HTML element can load from the counter server, but the Referer header and JavaScript page context may describe the embed iframe rather than the parent Wix page.Size the embed box to fit the counter image so it isn’t clipped or scrollable.
<img> tag.Starter sites are served on a webflow.io subdomain. Expect that as the referrer.
<img> inside an Embed code box is the likelier route.Check on the published site, signed out, whether the number changes between visits.
Official documentation: Google Sites: embed HTML and JavaScript
<img> tag pointing to a counter server is ordinary HTML, but confirm the external resource loads on your live page.The classic home of hand-made personal sites. Both counter types fit the culture.
<img> tag in HTML view.A sidebar gadget shows the counter on every page without editing each post.
img-src or script-src.On a host like this you can pick either type based on what you want to measure.
[img] works where images are allowed in signatures. Many forums (XenForo and Discourse, for example) can proxy or cache external images, which reduces counting accuracy.Read the forum’s signature rules. Size limits are common, and some communities ban counters outright.
03
Every counter answers one question: how many times did something reach the server? The approaches differ in what that “something” is and how much context comes with it.
The counter is an image URL. When the browser fetches it, the server adds one to the total and generates a picture of the current number, often in a retro digit style. This needs no code on the page and is the oldest approach on the web.
A small script runs in the visitor’s browser, collects context such as the page path, referrer and screen size, and sends it to the counter server. The script then displays the number as text or an image. More information, more requirements.
Your own web server or host logs every request it serves. This needs nothing on the page and isn’t affected by content blockers, but it counts bots too, and it’s only available where you can read the logs or the host provides statistics.
A hit is an HTTP request for a resource; one webpage can generate many hits. An image request is a hit for the counter image. A pageview is a page being viewed, which an image counter can only estimate from requests that reach its server. Unique visitors estimate distinct browsers or people over a defined period.
An estimate of how many distinct people or browsers visited. It’s calculated with cookies, browser storage, or by combining IP address and browser details over a time window. Shared networks, multiple devices and cleared cookies all distort it.
04
An HTML-only counter is just an image whose address points to a counter server instead of a file. The server counts the request and sends back a picture of the current number.
src points to a counter service and usually contains an ID for your counter.counter.example is a reserved placeholder domain. These snippets show the general shape of an image counter. They aren’t working code for any particular service, and don’t announce any service endpoint. Use the exact snippet your chosen counter provider gives you.
<!-- Illustrative example – not a real endpoint -->
<img src="https://counter.example/c/YOUR-ID.svg"
alt="Visitor counter"
width="88" height="31"
loading="eager"
referrerpolicy="no-referrer">[url=https://my-homepage.example/]
[img]https://counter.example/c/YOUR-ID.png[/img]
[/url]The example uses referrerpolicy="no-referrer" to avoid sending the page URL in the Referer header. Without that attribute, browsers generally send the site origin to an HTTPS counter on another origin, depending on site policy. The counter ID is fixed; no full page URL is added to the image address. Use loading="eager" if the counter should load immediately.
Counter servers send headers like Cache-Control: no-store so each view triggers a new request. Most browsers honor them, but some intermediaries don’t. The back button and “restore tabs” can also show the page without fetching the image again. Some services add a changing parameter to the URL, which only works when the page itself is generated per view.
To protect members’ IP addresses and avoid mixed-content warnings, many forums fetch external images through their own proxy. The proxy may fetch the image once and serve the same copy for hours or days. The counter then sees the proxy, not the readers, and may barely move.
Any external image reveals the visitor’s IP address, user-agent and often the referring origin to the server hosting it. That’s true of every embedded image, not just counters. Choose a provider whose privacy information you have read, and consider whether your site needs to mention the counter in its own privacy notice.
Image counters miss visitors who block images or third-party requests, and can be inflated by crawlers, link-preview bots and email prefetchers. They can’t record time on page, scroll depth or which link someone clicked. For a sense of popularity they’re fine; for decisions that matter, use more data.
05
Yes. An HTML image counter is a normal <img> tag whose source is a counter server. Each time a browser requests that image, the server records a hit and returns an image. It works wherever external images are allowed, but it counts image requests that reach the server, not every pageview.
On WordPress.com plans that don’t allow plugins or custom code, scripts are removed from posts and widgets, but external images can usually be inserted by URL or in a Custom HTML block. Whether the counter increments on each view needs testing, because WordPress.com may serve external images through its own image service, which caches them.
Often, if the forum allows [img] BBCode in signatures. Check the forum rules first: many limit image size or ban counters, and many forums route external images through an image proxy, which caches the image and hides the reader’s details from the counter server.
A hit is one HTTP request for a resource; one pageview can generate many hits. An image request is a hit for the counter graphic, not necessarily a pageview. A pageview means a page was viewed, and counters can only estimate this from the requests they receive. Unique visitors estimate distinct browsers or people during a defined period, using methods such as cookies or pseudonymous technical signals; all methods have limitations.
The counter only sees requests that reach its server. Browser caches, CDNs and forum or email image proxies can serve a stored copy instead; content blockers and data-saving modes can block external images; and some readers never load images at all. Bots and link previews can also add hits that aren’t real visits.
It depends on the service. Some counters set cookies or use browser storage to estimate unique visitors; others use none and count raw requests or hashed technical data. Read each provider’s privacy information, and check your own legal obligations, since whether a notice or consent is needed depends on what is collected and where you and your visitors are.
Usually, yes. Free static hosts such as Neocities and Netlify let you add either kind of counter. Free website-builder plans often block JavaScript, so an image counter or the platform’s own statistics may be the only options. Test on a published page, because editor previews behave differently.
06
Background reading on the web standards behind visitor counters, plus one example service.
Example service
If you want a ready-made counter, counter.gratis is a free visitor counter service you can add to a website with an embed snippet. Check its site for the embedding methods it currently supports and what it collects. This guide doesn’t assume an HTML-only image option is available there. Compare it with other providers before choosing.
Disclosure: Website Counter Guide is affiliated with counter.gratis. It’s mentioned here as one example, not as the only or best option. Nothing on this site is a paid placement for other services.
<img> element. Attributes like loading and referrerpolicy.Cache-Control affects repeat requests.[img] and [url].