Free SEO Tools That Actually Work in 2025

Most lists of free SEO tools are lists of trials. You click through, sign up, run two reports, and hit a wall that asks for a card. This is a different list. Every tool here is one you can use for real client work, month after month, without paying anything, and we tell you exactly where each one stops being useful.

The Three Kinds of Fake Free

Name the patterns and you stop wasting afternoons on tools that were never going to work.

The trial. Fourteen days of everything, then nothing. Fine for evaluating a purchase, useless inside a workflow: a tool that sits in your monthly reporting and then expires has not saved you money, it has delayed the invoice.

The freemium teaser. The tool works, but the free tier is calibrated to be just short of useful. Three keyword lookups a day. Ten rows of a four-hundred-row report. A legitimate business model, not a toolkit.

The lead magnet. A checker that audits one URL and emails a PDF so a sales team gets your address. Sometimes a useful number, never a dataset.

What follows is sorted by job, not vendor: what each tool does, its real limits, and where you would need to pay. At the end, a zero-dollar workflow that stitches four of them together.

Crawling and Auditing

Crawling is where free tiers hurt most, because URL caps land exactly where the interesting problems are. A 500-URL cap on a 12,000-page site does not give you a partial audit, it gives you a look at your navigation.

LibreCrawl

We build LibreCrawl, so treat this entry with suspicion and go read the repository. It is a desktop and self-hosted SEO crawler under the MIT licence, which means free in the sense that matters: no account, no seat count, no URL cap, no feature gating, and the source is there if you want to know what it does before pointing it at production.

What a crawl gives you:

• Unlimited URLs - the ceiling is your machine's RAM and your patience, not a licence tier.
• JavaScript rendering - a Playwright-backed mode for sites where content only exists after hydration. On React, Vue, Next, or Angular sites, crawling raw HTML tells you almost nothing, and the failure modes are subtle enough that we wrote a guide on crawling JavaScript websites.
• Issue detection - missing and duplicate titles, meta description problems, heading hierarchy, broken internal and external links, redirect chains, orphaned pages, thin content, images without alt text.
• Exports - full CSV and JSON dumps. A tool that shows problems but will not hand you the rows cannot be part of a repeatable process.

It is a crawler, not a rank tracker or a backlink index. It tells you what is wrong with your pages, not what your competitors' links look like. For the longer comparison against the commercial crawlers, including where they win, we keep a rundown of Screaming Frog alternatives.

Screaming Frog SEO Spider (Free Tier)

Screaming Frog is the tool most SEOs learned on, and its free tier deserves credit rather than a sneer. It is not a crippled demo: you get the real crawler, the real interface, and useful reporting, capped at 500 URLs per crawl.

For a brochure site, a local business, or a blog with a couple of hundred posts, 500 URLs is the whole site, and it is ideal for spot checks on one section. The cap becomes a wall on e-commerce, publishing, or anything with faceted navigation, and configuration is limited too: JavaScript rendering, custom extraction, scheduling, and API connections sit behind the licence.

Search Data

Here free tools are not merely adequate, they are better than anything you can buy. Third-party tools model search behaviour. Search engines report it.

Google Search Console

Irreplaceable, and the first thing to set up on any site. No paid tool has this data, because it comes from inside Google.

• Actual query data - the real searches that surfaced your pages, with impressions, clicks, CTR, and average position. Not estimates.
• Page-level performance - which URLs earn impressions, which have high impressions and terrible CTR (usually a title problem you can fix in an afternoon), which have quietly declined.
• Indexing reports - what Google indexed, what it excluded, and why. "Crawled, currently not indexed" and "Duplicate without user-selected canonical" are among the most useful diagnostic strings in SEO.
• URL Inspection - a live fetch showing the rendered HTML Google sees, the canonical it picked, and the structured data it parsed.
• Core Web Vitals - field data by URL pattern, from real Chrome users.

The limits: 16 months of Performance history, so export regularly. The interface samples and rounds on large sites, and query data is filtered for privacy, so clicks never reconcile perfectly with analytics. Positions are averages, so read them as trends. And it only covers Google, which is why the next entry matters.

Bing Webmaster Tools

Underrated, mostly because Bing's traffic share is small enough that people skip setup. That is a mistake. Bing feeds a meaningful amount of AI-assisted search, and it ships a site scanner Google does not offer at all: a real crawler that audits your pages for technical issues alongside query data, backlinks, and index coverage.

Its backlink report is more generous than Google's, and it shows links to sites you do not own, which makes it a free (if incomplete) competitor link source. Import your Search Console properties and setup takes minutes. Limits: smaller data volumes, and the scanner is nowhere near as thorough as a dedicated crawler. A second opinion, not an audit.

Page Speed and Core Web Vitals

PageSpeed Insights and Lighthouse

Two faces of the same engine, and the distinction matters. Lighthouse runs a lab test: a simulated load under fixed conditions, giving a repeatable score and a list of opportunities. PageSpeed Insights runs Lighthouse and adds field data from the Chrome User Experience Report, which is what real visitors experienced over the trailing 28 days.

Lab data answers "what should I fix", pointing at render-blocking resources, unoptimised images, oversized bundles, and missing image dimensions. Field data answers "does this matter", and it is what Google uses for the Core Web Vitals assessment. A page can score 68 in the lab and pass in the field, or score 95 and fail. The field number is the one that counts.

Lighthouse is free and local: built into DevTools, available as a CLI (npm install -g lighthouse), and scriptable across a list of URLs. PageSpeed Insights is free on the web, and its API is free too.

The gap is scale. Testing one page at a time in a browser tab stops working past about twenty URLs, which is exactly where people stop bothering. LibreCrawl integrates the PageSpeed Insights API into a crawl, so speed data arrives attached to the URLs you just crawled instead of in a spreadsheet you join by hand. Without a key you are on the low anonymous quota; add a free Google API key and it rises to 25,000 requests per day, enough to speed-test every page on most sites on a schedule.

CrUX and Chrome DevTools

The Chrome UX Report is queryable directly, so you can pull field data for any origin with enough traffic, including competitors'. And DevTools is a serious performance tool: the Performance panel, the throttled Network panel, and the Coverage tab (unused CSS and JavaScript) explain a bad score faster than any report.

Analytics

Google Analytics 4

Free at any volume most sites will reach. Sessions, engagement, conversions, attribution, and an event model you shape to your site. Turn on the Search Console integration so organic landing pages sit next to behaviour data.

The honest complaints: the event model is a real relearn after Universal Analytics, sampling and thresholding start earlier than people expect, and retention defaults to a shorter window than most want. Change retention on day one. For raw data, the BigQuery export is free to configure (you then pay BigQuery costs, cents for a small site).

Matomo and Plausible Community Edition

If your constraint is privacy or data ownership rather than budget, self-hosted open source is the answer. Both have paid cloud versions and a free self-hosted edition.

• Matomo - the closest open source equivalent to a full analytics suite. Visitor logs, goals, funnels, and SEO reports. Heavier to run: it wants PHP and MySQL, and it grows.
• Plausible Community Edition - deliberately minimal, cookieless, and fast. One dashboard that loads instantly and a script measured in kilobytes, running from a Docker Compose file.

Self-hosting is not free even when the licence is. You own the server, the upgrades, and the backups. For clients where sending analytics data to a third party is a problem, that trade is worth making.

Keyword and Content Research

The weakest category for free tools, and pretending otherwise helps nobody. Keyword data is expensive to produce, so the free options are either coarse or rationed.

Google Keyword Planner

Free with a Google Ads account, and you do not need a live campaign. It draws on Google's own data, which beats any scraped estimate at the source.

The caveat to internalise: without active ad spend, Keyword Planner buckets volumes into wide ranges instead of numbers. You see "1K - 10K" where a paid tool shows 3,400, and a tenfold range is not something you can prioritise with. It also groups close variants, flattening distinctions that matter for planning.

Use it for what survives the bucketing: finding terms you had not thought of, gauging relative competition, and confirming a topic has demand. Do not use it to argue one keyword beats another inside the same bucket.

Google Trends

Free, unlimited, and unique. Not volume but relative interest over time, by region, and between terms, which answers what volume cannot: rising or dying, seasonal and when the season starts, which phrasing wins, and where demand concentrates. The rising queries panel regularly surfaces terms no keyword tool has caught up with.

Question and SERP Research Tools

Tools in the AlsoAsked and AnswerThePublic family visualise the questions people ask around a topic, which helps when structuring content and FAQ sections. They are also the clearest freemium teaser here, so the limits without the spin: a handful of searches per day or month, results often truncated, and export usually the first thing behind the paywall. Enough for one article, not for a content programme.

Two free workarounds. Search Console's query report shows the long-tail questions you already rank for, which beats any generic list. And Google's own People Also Ask, related searches, and autocomplete are live and specific to the query you care about.

Structured Data

Schema Markup Validator

The validator at validator.schema.org succeeds Google's old Structured Data Testing Tool, and it checks markup against the Schema.org vocabulary rather than one engine's requirements. The right tool for whether your JSON-LD is correct: syntax, types, property names, nesting. It takes a URL or pasted code, so you can validate templates before they ship.

Google Rich Results Test

Different question. The Rich Results Test tells you whether Google considers your markup eligible for a specific rich result, which is stricter than being valid. It renders the page (including JavaScript, which matters for client-side schema), lists detected item types, and separates eligibility-blocking errors from warnings.

Run both. Valid but ineligible is common, usually a missing required property, and eligible is not the same as displayed, since Google decides per query. Structured data is one of the ten categories in our technical SEO audit checklist, worth working through alongside this list.

Backlinks

The hardest data to get free, because the cost is continuously crawling the web and storing the graph. Two options are worth your time.

Ahrefs Webmaster Tools

Ahrefs gives free backlink and site audit data for sites you verify ownership of, which is generous from a company whose index is its product. Verify through Search Console or DNS and you get referring domains, your backlink profile, broken inbound links worth reclaiming, and a site audit crawl. A real slice of a real index, not a teaser.

The boundary is precise: sites you own only. Competitor analysis, the most common reason people want backlink data, is not included at any meaningful depth.

The Search Console Links Report

Google's own view of who links to you: top linking sites, top linked pages, and anchor text distribution. Sampled and lagging, so counts will not match third-party tools. Its advantage is provenance, which makes it the primary source for spotting an influx of low-quality links or confirming a placement was picked up.

Utilities That Punch Above Their Weight

None of these are products. All of them solve problems people buy tools for.

• curl - the fastest way to see what a server returns. curl -I https://example.com/page gives status codes and headers, including X-Robots-Tag and canonical headers that never appear in the HTML. Add -L to reveal a redirect chain hop by hop, and -A to send a specific user agent.
• view-source - the cheapest rendering check. Compare view-source: against the DevTools Elements panel: the first is what the server sent, the second is the DOM after JavaScript. Copy that exists only in the second is a rendering problem.
• Robots.txt testers - Bing Webmaster Tools tests specific URLs against specific user agents before you deploy a rule that quietly deindexes a section.
• Hreflang generators - free generators build correct tag sets from URLs and locales, eliminating wrong region codes, missing return links, and absent x-default. A crawl afterwards confirms the tags are reciprocal.
• Sitemap and log tooling - sitemaps are XML any script can parse, and access logs are the only record of what bots really did. Neither requires a licence.

What You Still Need Paid Tools For

A free toolkit covers most of technical SEO and much of content work. Knowing the gaps separates a workable stack from wishful thinking.

• Rank tracking at scale - hundreds of keywords daily across locations and devices, with history and alerts, needs infrastructure that costs money. Search Console's average position is a trend line, not a rank tracker.
• Competitor backlink data - the reason link indexes are expensive. No free path to another site's referring domains at depth.
• Competitor keyword and traffic estimates - what a rival ranks for and roughly what it earns them.
• Precise search volume - Keyword Planner's buckets are the free ceiling.
• Team workflow - shared dashboards, scheduled client reports, seats, and alerting.

Free tools diagnose extremely well and handle competitive intelligence poorly. If your work is fixing sites, you can go very far at zero cost. If it is deciding which market to attack next, buy one subscription and keep the rest free.

A Complete Zero-Dollar Audit Workflow

Here is how the pieces fit into an end-to-end audit you could run for a client this week without a single invoice. Budget half a day for a mid-sized site.

Step 1: Establish What the Search Engine Sees (30 minutes)

Open Search Console before crawling anything, because it frames every other finding. In order:

1. The Pages report. Compare the indexed count to what you expect; a large gap either way is your headline finding.
2. Every exclusion reason with a meaningful count, written down with its number.
3. The Performance report, last 3 months against the previous period. Export the top 1,000 queries and top pages.
4. Core Web Vitals, mobile and desktop, noting which URL groups fail.
5. Sitemaps, submitted counts against discovered counts.
6. Import the property into Bing Webmaster Tools and start its site scan in the background.

Step 2: Crawl the Site (1 to 3 hours, mostly unattended)

Point LibreCrawl at the domain with no URL limit. Turn on JavaScript rendering if the site is framework-built, with a 5 to 7 second timeout. While it runs, watch the status code distribution: a rising 404 or 500 count early means a pattern worth investigating now rather than at the end.

Export the full dataset to CSV, then pull the lists you will report on: missing and duplicate titles and descriptions, pages with no H1 or several, thin pages under about 200 words, broken links, redirect chains, orphaned pages, and images without alt text.

Step 3: Reconcile the Crawl Against Search Console (45 minutes)

Most audits skip this step, and it is where the real findings live. Join your crawl export to your Search Console page export on URL, then read the mismatches:

• Crawled URLs with zero impressions. Either not indexed, or indexed and unwanted.
• URLs with impressions your crawl never found. Orphans, or pages reachable only by routes the crawl missed, and often valuable.
• High impressions with low CTR. Cross-reference title lengths from the crawl. Usually the fastest win in the audit.
• Anything Search Console calls a duplicate. Compare the canonical you recorded against the one Google chose.

Step 4: Layer In Speed Data (30 minutes)

Create a free Google API key for the PageSpeed Insights API and add it to LibreCrawl so speed data returns attached to the crawled URLs. The 25,000 requests per day quota covers a whole site rather than a sample. For the worst offenders, open Lighthouse in DevTools: the crawl says which pages are slow, Lighthouse says why. Then compare against the field data from step 1 and fix the pages that are slow in the lab and failing in the field first.

Step 5: Validate Structured Data and Templates (30 minutes)

You do not need to test every page, only every template. Pick one representative URL per template (home, category, product or article, author, contact) and run each through the Schema Markup Validator for correctness and the Rich Results Test for eligibility. Check the same set with curl -I for header-level surprises, and compare view-source to the rendered DOM on one JavaScript-heavy page. Finish with Search Console's enhancement reports to see which issues are site-wide.

Step 6: Prioritise and Write It Up (1 hour)

Sort findings by impact and effort, not by category. Critical blocks indexing or breaks pages: robots.txt errors, widespread 404s and 500s, a stray noindex, failing Core Web Vitals on money pages. High costs traffic you already earn: duplicate titles, missing descriptions on high-impression pages, redirect chains, orphaned pages with impressions. Everything else waits, and saying so explicitly is part of the value you deliver.

Then set a recurrence: re-crawl monthly and diff against the previous export, check Search Console weekly, and re-run the full audit quarterly.

Conclusion

The free SEO stack in 2025 is stronger than most people assume, largely because the two most valuable data sources (Search Console and Bing Webmaster Tools) come from the search engines themselves and cannot be bought, and because open source has closed the gap on crawling. You can audit a site end to end, prioritise every technical issue, measure real user performance, validate structured data, and monitor all of it without paying anything.

What you cannot get free is competitive intelligence. Rank tracking at scale, competitor backlinks, and rival traffic estimates rest on expensive infrastructure, and the free tiers claiming to offer them are teasers. Know that before you promise a client something your stack cannot deliver. If you need one of those things, buy exactly that one thing and keep the rest free.

Key takeaways:

• Learn to spot the three fake-free patterns: the trial, the freemium teaser, and the lead magnet
• Search Console and Bing Webmaster Tools are free, irreplaceable, and should be set up first
• Use an unlimited open source crawler so URL caps stop shaping your audits
• Read lab and field speed data as answers to different questions, and a free API key lifts the PageSpeed quota to 25,000 requests per day
• Budget for competitive data (rank tracking, competitor backlinks) and nothing else