Most SEO tools hand you a score and expect you to trust it. Seodar shows the arithmetic: 100 minus weighted penalties, capped per category, with all 195 checks named in public. It crawls after JavaScript runs, so it measures the page a visitor actually gets, and it says what it could not reach instead of quietly scoring it anyway. Accessibility is in the same crawl, not a separate product: axe-core against WCAG 2.2, every finding citing its criterion. Free scan of 25 pages, no account.
Maker
šHi Product Hunt,
I have spent sixteen years building systems that measure things, starting in a Samsung laboratory and spending the last nine in iGaming, where being wrong about a number has a price attached to it.
Seodar came out of a frustration I could not let go of. Every SEO tool I used handed me a score out of 100 and no way to check it. Why 61 and not 74? Which finding moved it? If I fix this one thing, what happens? You cannot answer any of those, so the number is decoration.
So this one shows its working. The score is 100 minus weighted penalties, capped per category, and you can see which check took which points. All 195 checks are listed publicly at seodar.io/checks with what each one costs you. When the crawler cannot reach something, it says so rather than scoring the page as if it had.
Accessibility went into the same crawl because it is the same fetch. Running axe-core against WCAG 2.2 while the page is already open is nearly free, and it turns out the people who need one usually need both.
It is free to try on 25 pages with no account. Paste a URL and you get a score and the findings.
What I would most like feedback on: does the score read as honest to you, or does it still feel like a number you are being asked to trust? That is the whole thesis and I would rather hear it now than in six months.
Petar
Report
@pekipekiCongrats on launching! Been testing it on a couple of agency sites this week. Search Console integration plus scheduled monitoring means I get pinged when a score drops instead of hearing it from the client. One question: does the API/MCP expose the per-check penalties too, or just the top-level score? Would love to pipe that into our internal dashboard.
Report
Maker
@pirkeĀ Thanks ā and that's exactly the loop it's meant to close. Hearing about a
drop from the client first is the worst version of that.
Both, and the numbers reconcile.
GET /v1/crawls/{id} returns summary.category_penalties ā the exact penalty
each of the six categories contributed. On a crawl I just pulled: 0.0, 0.02,
0.15, 0.27, 2.71, 0.2 ā 3.35 total ā health_score 97. So you can rebuild the
score in your dashboard rather than trusting the number.
Per check, GET /v1/crawls/{id}/findings gives every finding its own row with
check_id, severity, status and impact_score, paginated with limit/offset and
a total. summary.top_fixes carries the same check_id plus priority, affected
page count and sample URLs, if you want the ranked shortlist without paging
1,000+ findings.
The rest of what you'd want for a dashboard:
ā¢ /crawls/{id}/accessibility ā WCAG 2.2 by success criterion, with the axe
version, criteria_covered vs criteria_testable, and an "uncovered" list of
criteria a machine can't decide
ā¢ /crawls/{id}/pages ā every URL with status_code, depth, indexable,
word_count, inbound_internal, load_ms and finding count
ā¢ /crawls/{id}/segments ā per-segment scores plus the match_type and pattern
ā¢ /sites/{id}/rankings and /sites/{id}/search ā positions with device, and
Search Console clicks/impressions/CTR with top queries and pages
All read-only. Write methods aren't gated behind a permission check ā the
routes don't exist, POST/PUT/PATCH/DELETE return 405. Nothing spends a credit.
Rather than polling, point a webhook at it: the score_drop event is signed
HMAC-SHA256 over timestamp + "." + body, so you pull the moment it falls.
MCP is at https://api.seodar.io/mcp ā OAuth 2.1, PKCE S256, dynamic client
registration, scope seodar:read.
Report
Marko ā thank you for taking the time to write this properly.
The segments point is the one I'd underline. Scoring a blog and a shop as one number hides the thing you actually need to know, and most tools average it away. Glad it showed you where the drag was.
On the re-scan trigger: there's a usable version of that today. Seodar has an API, so a Shopify webhook or a WP hook on publish can kick off a scan without waiting for a plugin from me. Email me and I'll help you wire it up. A native plugin is a bigger commitment and I'd rather not promise a date I can't keep.
On keyword research ā you're right, it is newer than the audit, and it shows. Volume data and country filters are the two gaps I hear most, and they're the right two to name.
On Ahrefs, I'd rather not argue with it. Paying for a backlink database you barely open is a real thing, but so is needing that database when your job is different from yours. Being the tool someone reaches for because it answers one question well is the position I want.