Internet Archive Scraper: Search Items, Metadata and Wayback Snapshots

Search archive.org, fetch item metadata with file lists and download links, or list Wayback Machine snapshots of any URL, using only the Internet Archive's official public APIs. One clean dataset row per item or snapshot. Metadata and links only, no bulk downloading.

What you get

• Search mode: items matching a query, with title, media type, creator, date, description, subjects, collections, language, licence URL, download count, size and the item page link.
• Details mode: full metadata for identifiers you already know, plus the file list (name, format, size, MD5, SHA-1, modified time and a direct download URL per file).
• Snapshots mode: Wayback Machine captures of a URL, domain or path prefix: capture time, HTTP status, MIME type, content digest, size and ready-made snapshot links (normal and raw id_).

Use cases

• Build a list of public-domain books, films, audio or software on a topic, with licence and download counts.
• Look up the files in a known item and get checksums and links for your own, permitted download.
• Track how a web page changed over time (collapse identical captures to see only the changes).
• Find the first and last archived capture of a URL, or the available captures in a date range, for research, journalism or due diligence.
• Feed an AI agent or research assistant citable archive.org links (callable through Apify's MCP server, https://mcp.apify.com).

What it does

Official endpoints only, no HTML scraping, no login:

• Search: https://archive.org/advancedsearch.php?q=...&fl[]=...&rows=&page=&output=json
• Details: https://archive.org/metadata/{identifier}
• Snapshots: https://web.archive.org/cdx/search/cdx?url=...&output=json&from=&to=&limit=&filter=&collapse=digest

Polite by design: sequential requests at most 1 per second, an identifying User-Agent, retries with exponential backoff on HTTP 429 and 5xx (honouring Retry-After), and a clear failure message otherwise. It never uses proxies or workarounds. It does not download item files: respect the Internet Archive's terms of use and each item's licence or rights statement when you use the links it returns.

Input

Field	Meaning
mode	search (default), details or snapshots
query	Search query, Lucene syntax accepted (creator:nasa AND subject:moon)
mediatype, collection	Search filters (texts, movies, audio, software, image, data, web, ...)
dateFrom, dateTo	Item date range in search mode; capture date range in snapshots mode (YYYY, YYYY-MM or YYYY-MM-DD)
sort	Relevance (default), most downloaded, newest or oldest, recently added
maxItems	1-10,000 items in search mode (default 50)
identifiers	Details mode: identifiers or archive.org/details/... URLs (up to 500)
includeFiles, includeDerivatives	Details mode: file list on/off; derivative files off by default
urls, matchType	Snapshots mode: URLs or domains (up to 100); exact, prefix, host, domain
collapseByDigest, statusCodeFilter, mimeTypeFilter, maxSnapshots	Snapshots mode: skip unchanged captures (default on), keep only status 200 (default), MIME type, 1-10,000 captures per URL (default 100)

Example input and output

Input (bundled test-input.json): {"mode": "search", "query": "apollo 11 mission audio", "mediatype": "audio", "dateFrom": "1969", "dateTo": "1970", "maxItems": 5}.

The records below use illustrative values only; real results depend on what archive.org holds when you run it. Our tests use synthetic fixtures shaped like the documented API responses.

{
  "kind": "item", "identifier": "example-apollo-audio", "title": "Example Apollo audio",
  "mediatype": "audio", "creator": ["NASA"], "date": "1969-07-20T00:00:00Z", "downloads": 1234,
  "collection": ["nasa"], "license_url": null,
  "item_url": "https://archive.org/details/example-apollo-audio",
  "metadata_url": "https://archive.org/metadata/example-apollo-audio"
}

{
  "kind": "snapshot", "queried_url": "example.com", "original_url": "http://example.com/",
  "timestamp": "20200101120000", "captured_at": "2020-01-01T12:00:00Z", "status_code": "200",
  "mime_type": "text/html", "digest": "SAMPLEDIGEST", "length_bytes": 1234,
  "snapshot_url": "https://web.archive.org/web/20200101120000/http://example.com/"
}

Limitations (read before relying on it)

• Metadata is uploader-supplied and often incomplete or inconsistent (dates in several formats, missing creators, free-text licences). Check rights before reuse; a link is not a licence.
• Advanced Search returns at most about 10,000 results per query; narrow the query with filters or dates for more. Counts and ranking are archive.org's.
• The Wayback CDX index lists captures it holds; it is not a guarantee that a page was captured on a given day, and recent captures can take time to appear. limit applies per URL after filtering and collapsing.
• Items that are dark, restricted or removed return limited or no metadata. Identifiers that are not found are skipped (listed in the run's SUMMARY record and the log), not billed.
• Details mode lists files of an item but never fetches them. Large items can have thousands of files per row.
• Archive.org can rate limit or be slow; the actor retries and then fails with a clear message rather than hammering the service. Not affiliated with or endorsed by the Internet Archive.

Pricing (pay per result)

Pay-per-event with Apify's built-in dataset-item event: each item (search, details) or snapshot is one result. The live price is set in Apify Console and shown on the Store page. Planning figure in our notes: USD 1.00 per 1,000 results (not yet published).

FAQ

Does it need an API key? No, the APIs are public and keyless.

Can it download the files? No. It returns metadata and links only, to respect the Internet Archive's terms and avoid bulk downloading.

How do I see only the changes of a page over time? Use snapshots mode with collapseByDigest on (default); identical consecutive captures are skipped.

How do I get the snapshot as it was archived, without the Wayback toolbar? Use raw_snapshot_url (the id_ form).

Can I search by creator or subject? Yes, put Lucene syntax in query, for example creator:"Library of Congress" AND subject:maps.

Why did an identifier return nothing? It does not exist, was removed, or is dark. See the SUMMARY record.