Free SEO tools for a small project

Running SEO on a small project does not require a hundred-euro-a-month suite. It requires seven pieces: knowing what Google sees of you, what people do once they arrive, whether the site loads well, whether the structured data is valid, whether links are broken, whether the headers are set correctly, and what people actually search for. All of them have a genuinely free option.

Here is the list, in the order worth setting them up. Each service’s concrete limits are on its page in the directory, because they change often.

1. Search Console, before anything else

It is free, it is Google’s, and it is the only source that tells you what Google sees of your site: which pages it indexed, which it discarded and why, which queries bring you impressions, and your average position on each.

No paid tool replaces this, because the rest estimate and this one measures. Verify it on the project’s first day: the data isn’t retroactive, so every week you delay is a week you will never have.

The report that pays most isn’t performance, it is coverage: excluded pages appear there with the reason written out. “Duplicate, Google chose a different canonical” is the warning that something from point 3 of technical SEO for developers is wrong.

Add Bing Webmaster Tools too. It takes ten minutes, and since several AI assistants lean on the Bing index, it matters more than its market share suggests.

2. Analytics that doesn’t force a banner

Search Console tells you how people arrive; analytics tells you what they do once there. For a small project the best return is cookieless analytics: less friction, no consent banner and vastly less legal work.

• Umami — cookieless, with an open source version you can self-host whenever you want.
• Cloudflare Web Analytics — cookieless and without a heavy script if you are already behind Cloudflare.
• Microsoft Clarity — heatmaps and session recordings, useful for understanding why a page doesn’t convert.
• Ahrefs Web Analytics — free analytics from a company that knows SEO.

On the free plan, what to check is the monthly event cap and retention: keeping thirty days is fine for tuning, not for comparing against last quarter. Full category in analytics.

3. Lighthouse and PageSpeed Insights

Lighthouse is already in your browser, inside the developer tools. PageSpeed Insights is the same thing on the web, and additionally shows real-user data once your site has enough traffic.

Two warnings so you don’t waste time:

• Lighthouse’s SEO score is shallow. It checks there is a title, a description, and that links are crawlable. A 100 means nothing more than that.
• Lab data varies a lot between runs. If you are going to compare before and after, use real-user data, not a one-off measurement on your laptop.

What is genuinely valuable is the opportunities list: almost always unsized images and render-blocking resources.

4. The Rich Results Test

Google’s tool for validating structured data. You give it a URL or an HTML snippet and it tells you whether the markup is valid and whether it is eligible for any rich format.

Use it once per page type, not per page: if your article template is right, all hundred articles are right. And back it up with an automated check that the JSON at least parses, which is the silly failure nobody sees. The rest of the subject is in structured data: what to mark up and what not.

5. A crawler, even if it is your own

You need something that walks the site and tells you what is broken. The options:

• A desktop crawler with a free tier limited to a few hundred URLs. For a small site, plenty.
• Firecrawl, if you would rather crawl via API and process the result yourself.
• Your own script, which is what we ended up doing.

The third is the one I would actually recommend, and not for budget reasons: your own script checks your rules, not generic ones. Ours verifies that the sitemap covers everything public, that no public page carries noindex, that titles are in range, that hreflang is complete, that the JSON-LD parses and that there are no broken internal links. It exits with code 1 and hangs off CI.

It found things that had been broken for weeks with nothing warning us. It is the best return on this list and the only one that doesn’t exist until you write it.

6. Header and security checks

Not SEO in the strict sense, but a site with badly configured HTTPS or missing headers sends bad signals and invites trouble.

Mozilla Observatory analyses your security headers for free and gives you the list of what’s missing along with why it matters. Ten minutes, once, and it is settled. More in security.

7. Keyword research, without spending

This is where the expensive tools live, and there is a fair amount you can do without them:

• Search Console itself. Queries where you already sit in positions 8–15 have the best return: they already find you, they just need a push. No external tool has this.
• Search autocomplete and related questions. Free, and literally what people type.
• The limited free tiers of the well-known suites, which give you a handful of queries a day. Enough to validate an article idea.

One piece of advice worth more than any tool: before writing, check that no page of yours already competes for that exact phrase. When two pages of the same site fight over one keyword, Google splits the authority and neither ranks well.

The order if you are starting today

Search Console and analytics on day one, because the data isn’t retroactive. The check script in week one, because it prevents the silent damage. Everything else, once there is something to measure.

Frequently asked questions

Can you do SEO without paying for tools? For a small project, yes, and rather well. Search Console covers what no paid tool can give you, and the technical side automates with your own script. Paid suites start paying off when you need competitor analysis and link profiles at scale.

Do I need Google Analytics? No. It is powerful and free up to a point, but it requires a consent banner and compliance work. To know which pages work, cookieless analytics gives the same answer at far lower cost.

How often should technical SEO be reviewed? Don’t review: automate. A quarterly manual review finds the problem three months late; a script in CI finds it in the commit that introduced it. Save manual review for what needs judgement, which is the content.

Which free tool replaces the paid suites? None, and it is worth being honest. What the suites do that you cannot replicate is competitor analysis and link profiles, which need their own index of the web. For your own site, though, you have practically everything.

Are online SEO graders with a score useful? Not very. They score superficial things and award points for elements that no longer matter. As an initial checklist they can help; as a way to prioritise work, no.