Disclosure: I'm Lindenbyte, the small studio that built the Apify Actor used below, and this post was
drafted with AI assistance. All numbers come from real runs on 2026-11-16.
Most "free SEO checkers" look at one URL at a time. The problems that actually hurt a site (missing meta
descriptions on 40 blog posts, the same title on every category page, a page with no H1) only show up when
you crawl the whole thing. Desktop crawlers do this well, but they are tied to one machine and hard to
schedule. Here is a lightweight alternative you can run from a browser tab or a Python script.
1. Crawl a site
Open the SEO Audit Crawler and give it
a domain. I started with ghost.org and capped the crawl at 25 pages:
{
"startUrls": ["ghost.org"],
"maxPagesPerSite": 25,
"maxDepth": 3
}
It follows same-site links breadth-first, respects robots.txt, and audits each HTML page with plain HTTP
(no headless browser). The run took 4.4 seconds and cost a fraction of a cent in compute.
2. Read the per-site summary
The first row of the dataset is a free summary per site. For ghost.org:
- 25 pages crawled, 20 with at least one issue, average page score 97/100
- 0 errors, 6 warnings, 19 notices
- most common issues:
images-missing-alt(12 pages),title-too-short(4),h1-multiple(3),h1-missing(2) - no duplicate titles, meta descriptions or H1s
Then one row per page with the issue codes, title and meta description (with lengths), H1 count, canonical,
robots meta, word count, images missing alt text, hreflang, broken internal links and response time. Sorting
by score puts the worst pages first: here /help scored 86 (17-character title, short meta description,
no H1).
3. Compare yourself with competitors
Put several sites in one run. I crawled 20 pages each of three programming-language sites:
{ "startUrls": ["python.org", "ruby-lang.org", "go.dev"], "maxPagesPerSite": 20, "maxDepth": 2 }
| Site | Avg score | Warnings | Top issues |
|---|---|---|---|
| python.org | 89 | 7 | short meta descriptions (18 pages), duplicate meta description (18), no canonical (20) |
| go.dev | 88 | 27 | images without alt (20), no meta description (19), no Open Graph tags (20) |
| ruby-lang.org | 84 | 38 | short titles (19), images without alt (19), no H1 (14) |
The 60 pages took one run. A table like this is a decent opener for a client pitch or a quarterly SEO review.
4. Pull it into Python
from apify_client import ApifyClient
client = ApifyClient("<APIFY_TOKEN>")
run = client.actor("lindenbyte/seo-audit-crawler-bulk-site-checker").call(
run_input={"startUrls": ["example.com"], "maxPagesPerSite": 50})
rows = list(client.dataset(run.default_dataset_id).iterate_items())
pages = sorted((r for r in rows if r["type"] == "page"), key=lambda r: r["score"])
for p in pages[:10]:
print(p["score"], p["url"], ", ".join(p["issues"]))
Schedule it weekly (Apify schedules or a cron job) and diff issueCount per URL to catch regressions after
a redesign or CMS update.
5. Limits, honestly
- No JavaScript rendering: a client-side-rendered app shows up with empty titles and H1s. Check one page manually if scores look suspiciously bad.
-
redirectedfires for every page reached through a redirecting link. python.org links to its ownhttp:/// non-www URLs, so all 20 pages got that notice. It is a hint to clean up internal links, not an emergency. - It does not measure speed or Core Web Vitals; pair it with a PageSpeed tool for that. It also does not check external links or backlinks.
- Free alternatives exist (Google Search Console for your own site, desktop crawlers with free page limits). What this adds: many sites per run, no install, API/JSON/CSV output, schedulable.
Pricing is pay per event: $0.004 per audited page plus a small start fee; error rows and site summaries are
free, and maxPagesPerSite caps your spend. Feedback and feature requests: the Issues tab on the Actor page.
Top comments (0)