Seo Jsonld Extractor avatar

Seo Jsonld Extractor

Pricing

$0.01 / actor start

Go to Apify Store
Seo Jsonld Extractor

Seo Jsonld Extractor

Pricing

$0.01 / actor start

Rating

0.0

(0)

Developer

Eichan Jolly

Eichan Jolly

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

4 days ago

Last modified

Categories

Features

Limited permissions

Share

JSON-LD Block Extractor

What it does

Fetches one page and extracts every HTML <script type="application/ld+json"> block in document order, returning one dataset item per block. It retains the exact parsed JSON value in data, and gathers the actual @type string values found anywhere in that value. Malformed JSON becomes its own item with a readable parseNote; later blocks are still processed.

Who it is for

SEO auditors, structured-data maintainers, site owners, and developers checking what JSON-LD a page actually embeds.

Input

Supply one publicly reachable HTTP/HTTPS URL (no account or API key):

{
"url": "https://yoast.com/author-publisher-entities-seo/"
}

This is also the prefilled Store example; it is a real public page. It may change after publication. The only configurable field is url. The Actor will request the target site’s /robots.txt and, if permitted, fetch the page itself.

Output

sourceUrl (string or null on invalid input), blockIndex (1-based integer or null), types (list of discovered @type strings), data (parsed JSON value or null), status (ok, malformed, no_jsonld, or error), parseNote (string or null), error (string or null).

Example actual dataset item, recorded by the included local script-mode smoke test (the 127.0.0.1 address is a disposable test fixture, not the public prefill):

{
"sourceUrl": "http://127.0.0.1:43373/page",
"blockIndex": 1,
"types": [
"Article",
"Person"
],
"data": {
"@context": "https://schema.org",
"@graph": [
{
"@type": "Article",
"headline": "Test post"
},
{
"@type": "Person",
"name": "Eichan"
}
]
},
"status": "ok",
"parseNote": null,
"error": null
}

The script-mode fixture includes one valid JSON-LD @graph and one malformed JSON-LD block, producing two records. See tests/evidence/smoke_dataset.json for all actual smoke-test records, including malformed JSON-LD when applicable.

How to run

On Apify, paste the URL in the input form and run the Actor. The input is loaded in this exact precedence order:

  1. Parse ACTOR_INPUT_JSON if set (including when it contains invalid JSON; no silent fallback).
  2. Otherwise use GET https://api.apify.com/v2/key-value-stores/{APIFY_DEFAULT_KEY_VALUE_STORE_ID}/records/INPUT with Authorization: Bearer {APIFY_TOKEN} when the default key-value store ID exists.
  3. Otherwise read ./storage/key_value_stores/default/INPUT.json relative to the process working directory.

On Apify, results are appended to the default dataset by an authorized REST API POST. Locally, one JSON file per item appears under storage/datasets/default/.

To run locally from this Actor folder:

export ACTOR_INPUT_JSON='{"url": "https://yoast.com/author-publisher-entities-seo/"}'
python src/main.py

Or put the same JSON in storage/key_value_stores/default/INPUT.json and run python src/main.py with ACTOR_INPUT_JSON unset. Only Python 3.12+ standard-library modules are required.

To rerun verifiable offline tests:

python -m unittest discover -s tests -p test_actor.py -v
python tests/run_smoke.py

The smoke runner launches a local HTTP fixture and executes python src/main.py as a subprocess with a valid ACTOR_INPUT_JSON. It writes the full terminal output, valid input, and produced dataset records to tests/evidence/. ACTOR_TEST_ALLOW_LOOPBACK=1 is only set by this local test runner, never by default.

Limits and error handling

It reads server-returned HTML only, without running JavaScript. JSON-LD added by client-side scripts is not included. It does not follow src attributes, external URLs, or validate schema.org semantics or Google rich-results eligibility. A page with zero JSON-LD blocks produces one no_jsonld item, rather than an empty dataset. Block indexes start at 1 and duplicate blocks are preserved. Malformed JSON is not repaired or inferred.

  • One page per run. No general crawling, search-result scraping, login, or third-party Python packages.
  • Two HTTP requests maximum in a normal run (robots.txt and the page), with at most three same-host redirects for each request. The page is limited to 2 MB; robots.txt is limited to 200 KB; each request has a 12-second timeout.
  • By default, local/private/reserved IP addresses and nonstandard ports are rejected; cross-host redirects are rejected. The actor checks DNS addresses but is not a hardened defence against DNS rebinding on hostile hostnames. Use only public sites you are authorized to inspect.
  • Robots.txt opt-outs are respected; missing robots.txt (404/410) is allowed. If the page cannot be fetched, or input is invalid, the Actor writes a stable error dataset record and prints a short message, not a traceback. A failed run may exit nonzero even if its error item is written.
  • Web content, robots policies, and site availability change. The public example is not part of the offline unit/smoke fixture, so cloud execution must be checked after deploying.

Suggested Store categories and pricing

Categories: SEO tools, Developer tools, Marketing.

Suggested $0.01 per run, independent of the number of blocks. This is a proposed Store price only; no monetization or billing is configured.

Deployment state

Source-only deliverable, not published. No Apify cloud runs or billing configuration have been performed.