Page audit

Before you look at rankings, look at the page. This endpoint fetches a URL the way a browser would and returns everything an on-page review needs: a score built from twelve rules you can read, the elements behind each rule, and the timing and crawl signals that decide whether Google indexes the page at all. It runs on our own servers, so it costs half a credit and answers in a couple of seconds.

POST/v1/pages/audit

Audit a URL

Returns the audit of one page. The score starts at 100 and loses the weight of every check that fails, so two pages with the same score have the same problems. warnings lists things that do not change the score but a human should know, like a noindex or a canonical pointing elsewhere.

Costs 0.6 credits per call. Pages are read up to 2 MB; larger ones are analysed on their first 2 MB and say so.

Required attributes

Name
url
Type
string
Description
The page to audit. https:// is assumed when you leave the scheme out, and http:// is upgraded. Private and internal addresses are refused.

Checks and weights

Check	Weight
Title present, 10 to 60 characters	12
Meta description, 50 to 160 characters	12
Exactly one H1	10
Served over HTTPS	10
Every image has alt text	10
Canonical URL declared	8
Open Graph tags present	8
Valid JSON-LD structured data	8
HTML under 200 KB	6
Response under 1500 ms	6
`lang` attribute on `<html>`	5
Twitter Card tags present	5

Response fields

Name
score
Type
integer
Description
0 to 100. Each failed check subtracts its weight.
Name
indexable
Type
boolean
Description
false when the page answers an error, carries noindex, or robots.txt disallows its path.
Name
checks[]
Type
array
Description
One entry per rule with id, label, weight, passed and a detail that says what was found.
Name
warnings[]
Type
array
Description
Findings outside the score: noindex, robots.txt blocks, missing viewport, thin content, canonical pointing to another URL.
Name
page
Type
object
Description
Title and description with their lengths, canonical, robots meta, lang, viewport, word count, headings by level, image and link counts with examples, Open Graph and Twitter tags, JSON-LD types, hreflang alternates.
Name
performance
Type
object
Description
Time to first byte, total time, HTML size, compression, server header, cache control and how many scripts, stylesheets, iframes and images the HTML references. Measured for the document only, without rendering.
Name
robots_txt
Type
object
Description
Whether robots.txt exists, the sitemaps it declares, and whether it disallows this path for all crawlers.

Request

POST

/v1/pages/audit

curl -X POST https://suvjexogmfpvxjcktvuk.supabase.co/functions/v1/api/v1/pages/audit \
  -H "Authorization: Bearer $PENGU_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://mailchimp.com/"}'

Response

{
  "data": {
    "url": "https://mailchimp.com/",
    "final_url": "https://mailchimp.com/",
    "status": 200,
    "fetched_at": "2026-09-10T07:56:59.897Z",
    "score": 72,
    "indexable": true,
    "checks": [
      { "id": "title", "label": "Title present, 10 to 60 characters", "weight": 12, "passed": true, "detail": "41 characters" },
      { "id": "meta_description", "label": "Meta description, 50 to 160 characters", "weight": 12, "passed": false, "detail": "216 characters" },
      { "id": "single_h1", "label": "Exactly one H1", "weight": 10, "passed": true, "detail": "1 H1 found" },
      { "id": "image_alt", "label": "Every image has alt text", "weight": 10, "passed": false, "detail": "2 of 94 images without alt" },
      { "id": "html_size", "label": "HTML under 200 KB", "weight": 6, "passed": false, "detail": "487.3 KB" },
      // ...
    ],
    "warnings": [],
    "page": {
      "title": "Email Marketing Platform | Mailchimp",
      "title_length": 41,
      "meta_description_length": 216,
      "canonical": "https://mailchimp.com/",
      "robots_meta": "index,follow,archive",
      "lang": "en",
      "word_count": 2395,
      "headings": {
        "h1": ["Email and SMS marketing without the learning curve"],
        "h2": ["Most popular", "Recommended for your business", "..."],
        "counts": { "h1": 1, "h2": 13, "h3": 16, "h4": 28, "h5": 0, "h6": 0 }
      },
      "images": { "total": 94, "missing_alt": 2, "lazy_loaded": 61, "missing_alt_examples": ["..."] },
      "links": { "internal": 212, "external": 9, "nofollow": 0, "external_examples": ["..."] },
      "structured_data": ["WebPage"],
      "hreflang_alternates": 12
    },
    "performance": {
      "ttfb_ms": 212,
      "total_ms": 640,
      "html_size_kb": 487.3,
      "compression": "br",
      "resources": { "scripts": 31, "stylesheets": 4, "iframes": 0, "images": 94 }
    },
    "robots_txt": { "present": true, "sitemaps": ["https://mailchimp.com/sitemap.xml"], "disallows_page": false }
  },
  "meta": {
    "credits_charged": 0.5,
    "credits_remaining": 483,
    "request_id": "6c1d0f7a-2f3b-4f0e-9d21-8a5c4b0e1f22"
  }
}

What it is good for

First call of any audit. Run it on the home page and the three pages that matter most before spending credits on rankings. Most quick wins are here.
Batch checks from an agent. Ask Claude to list a site's URLs with the sitemap endpoint below, audit the newest fifty and table the scores. At half a credit a page, a hundred pages cost fifty credits.
Before and after. Audit, fix, audit again. The checks are deterministic, so a score change means a real change.

Pages are fetched from our servers in Europe. A site that redirects visitors by location will do so here too, and final_url says where you landed; pass the exact localised URL when it matters, for example mailchimp.com/en-us/ instead of mailchimp.com.

It reads the HTML as served, without running JavaScript. A page whose title or content only exists after client-side rendering will look thin here, which is itself worth knowing: Google renders JavaScript, but later and less often.

POST/v1/pages/sitemap

Sitemap URLs

The URLs a site declares in its sitemaps, with their lastmod. Sitemaps are found through robots.txt and the usual locations (/sitemap.xml, /sitemap_index.xml), sitemap indexes are followed, and gzipped files are read. Up to 500 URLs come back, newest first when the site dates them.

Costs 0.6 credits per call.

Required attributes

Name
site
Type
string
Description
A domain such as hubspot.com, a site URL, or a direct sitemap URL when you already know it.

Optional attributes

Name
limit
Type
integer
Description
URLs to return, 10 to 500. Default 100.

Response fields

Name
sitemaps
Type
array
Description
The sitemap files that were read, in order.
Name
truncated
Type
boolean
Description
true when the site declares more URLs or child sitemaps than were read. Large sites usually do.
Name
urls[]
Type
array
Description
url and lastmod (ISO date or null).

Request

POST

/v1/pages/sitemap

curl -X POST https://suvjexogmfpvxjcktvuk.supabase.co/functions/v1/api/v1/pages/sitemap \
  -H "Authorization: Bearer $PENGU_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"site": "hubspot.com", "limit": 50}'

Response

{
  "data": {
    "site": "https://www.hubspot.com",
    "sitemaps": [
      "https://www.hubspot.com/sitemap.xml",
      "https://www.hubspot.com/sitemap-products.xml"
    ],
    "count": 50,
    "truncated": true,
    "urls": [
      { "url": "https://www.hubspot.com/products/crm", "lastmod": "2026-09-08" },
      { "url": "https://www.hubspot.com/products/marketing", "lastmod": "2026-09-08" },
      // ...
    ]
  },
  "meta": {
    "credits_charged": 0.5,
    "credits_remaining": 482.5,
    "request_id": "4a7f2c19-8e3d-4b6a-9f10-2c5d7e8f9a01"
  }
}