Internet Archive Scraper | Search, Files & Reviews

Search Archive.org and export public catalogue items, item metadata, reviews, ratings or file lists to CSV, Excel and JSON. Use it for library catalogues, research datasets, digital collections and finding public download links.

Per 1,000 delivered rows: $0.50 for items, $0.20 for reviews, or $0.10 for files. No startup fee. No Archive.org account or API key is required.

Start in three steps

1. Choose Search catalogue, Get item details, Export reviews or List files.
2. Enter URLs, identifiers or search queries: use item URLs/identifiers for details, reviews and files, or queries for Search catalogue.
3. Set Maximum results, add a spending limit and click Start. Open Results or download the dataset in your preferred format.

The form has only three controls: Operation, URLs, identifiers or search queries, and Maximum results. The default example collects one item at the supported 1 GB memory setting. Technical collection allowances are sized automatically within safety caps.

What you can collect

Operation	Results
Search catalogue	Matching item identifiers, titles, links and creators when shown on the search cards. Supports books, audio, video, images and software.
Get item details	Available title, creators, description, publication information, publisher, language, subjects, collections and public statistics.
Export reviews	Public review text, title, star rating, reviewer, displayed date and a link to the source review.
List files	File name, extension, public download link, displayed file size and modification date. Files themselves are not downloaded.

Only fields displayed by Archive.org are collected. Item views are kept separate from download counts. Missing ratings remain unknown, and rounded file sizes stay as displayed text.

Use cases

• Library catalogues: collect available titles, creators, publication details and item links for spreadsheet joins.
• Research datasets: find public books, audio, video, images and software using Archive.org search expressions.
• Review analysis: export review text and ratings, filter the exported dataset in your own tools, and inspect summaries of the delivered sample.
• File inventories: list visible filenames, displayed sizes and public download links without downloading the files.

Input examples

Get item details

{
  "operation": "itemDetails",
  "targets": ["https://archive.org/details/goodytwoshoes00newyiala"],
  "maxResults": 1
}

Search Archive.org

{
  "operation": "search",
  "targets": ["title:\"Goody Two-Shoes\""],
  "maxResults": 10
}

Use keywords or Archive.org search expressions such as creator:(Charles Dickens), subject:astronomy, collection:librivoxaudio, mediatype:texts, language:eng and date:[1900 TO 1950]. Combine expressions using AND or OR. For Search catalogue, enter your queries in the shared input instead of the example item URL. Search expressions retain their normal Archive.org meaning.

Export reviews and ratings

{
  "operation": "reviews",
  "targets": ["goodytwoshoes00newyiala"],
  "maxResults": 10
}

Reviews are exported without text or star filtering, including reviews with unknown or zero ratings. Filter the exported dataset afterwards. Old API inputs minStars, maxStars and reviewTextContains are accepted but ignored, with a warning in the log and run summary.

The Review summary includes rating distributions, averages for known ratings, recurring terms and question-like excerpts from delivered reviews. The optional JSON/API field analysisQuery adds literal phrase matches to the summary; it never suppresses exported rows or uses AI models.

List file names and download links

{
  "operation": "files",
  "targets": ["goodytwoshoes00newyiala"],
  "maxResults": 10
}

Batch several item URLs or identifiers in targets. Repeated inputs and repeated results are deduplicated within a run. Each operation returns one kind of row, making spreadsheet exports straightforward.

Output

The default dataset contains only requested result rows. For example, a file result has this shape; illustrative values are not a current source snapshot:

{
  "entityType": "file",
  "identifier": "example_item",
  "fileName": "example_item.pdf",
  "fileExtension": "pdf",
  "fileSizeText": "1.4M",
  "downloadUrl": "https://archive.org/download/example_item/example_item.pdf",
  "sourceUrl": "https://archive.org/download/example_item",
  "collectionSource": "rendered_public_directory",
  "dataInputSource": "web",
  "sourceVerified": true
}

Use the Items, Reviews or Files dataset view for the selected operation. Public page coverage and delivery reports pages loaded, confirmed result counts, source failures and stopping reasons. Errors and summaries are not billed as results. Check the summary when a run returns fewer rows than requested.

Pricing and limits

Delivered rows	Catalogue items	Reviews	Files
1	$0.0005	$0.0002	$0.0001
100	$0.05	$0.02	$0.01
1,000	$0.50	$0.20	$0.10

These rates apply on every Apify plan. Search catalogue and Get item details charge once per delivered item. Export reviews charges once per delivered review; List files charges once per delivered file. Each row is charged only for its own type, with no additional item fee for review or file exports. Supplied-review processing uses the review rate. Actor execution costs are included; there is no startup or page fee. A run with no delivered rows has no Actor result charge. Standard Apify storage access after a run can carry separate platform costs.

Maximum results defaults to 1 and supports up to 1,000 per run. Requests using targets automatically allow up to Maximum results reviews per item. Their page allowance is the larger of three, Maximum results, or the number of targets, capped at 100. Opening pages and visible pagination or scrolling share this allowance. JSON/API clients can explicitly set maxPublicPages or maxReviewsPerItem within their existing ranges; requests without targets retain defaults of three pages and ten reviews per item. Collection stops after 240 seconds to leave time for delivery and shutdown. Page, time, result and spending limits can produce a partial sample.

The Actor reads rendered public pages in a browser. It does not call source APIs, parse intercepted API responses, log in, bypass restricted downloads or download file contents. Search scrolling and review expansion depend on the controls currently exposed by Archive.org. File lists contain visible directory entries, not a complete API manifest with checksums. Wayback snapshots and live change monitoring are not included.

Python API example

import os
from decimal import Decimal
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])
run = client.actor("thescrapelab/internet-archive-book-reviews-scraper").call(
    run_input={
        "operation": "reviews",
        "targets": ["goodytwoshoes00newyiala"],
        "maxResults": 10,
    },
    memory_mbytes=1024,
    max_total_charge_usd=Decimal("0.05"),
)
if run is None:
    raise RuntimeError("The Actor did not return run information.")
dataset_id = run.default_dataset_id if hasattr(run, "default_dataset_id") else run["defaultDatasetId"]
for row in client.dataset(dataset_id).iterate_items():
    print(row)

The dataset API also works with Apify integrations such as Make, n8n, Zapier and Google Sheets. Use the Actor's API tab for your preferred client.

Existing JSON/API integrations

API calls without targets retain their existing source precedence: startUrls, then identifiers, then sources. Search continues to accept searchQueries. API calls omitting both operation and targets retain the previous public-review or supplied-record workflow.

When supplied, targets takes precedence over older source/query fields and requires an explicit operation. Empty targets are rejected. Existing field names remain accepted; inactive legacy settings do not enable unsupported features. The two hidden collection allowances now use runtime defaults so that new shared-input requests can size them automatically.

Migration for review filters: minStars, maxStars and nonblank reviewTextContains no longer filter any collection path. Their names appear in ignoredFilters, with a readable warnings message in the applicable run summary and logs. Remove these settings and filter the resulting CSV, Excel, JSON or dataset in your own workflow.

JSON/API supplied-record processing still accepts providedRecords or providedDatasetId, validates review records and removes duplicates without revisiting their websites. Non-empty supplied records or a supplied dataset take precedence over web inputs. Clear the JSON array to [] for web collection. Supplied records remain sourceVerified: false; this path can use 256 MB when explicitly selected. Its record limit is maxItems.

Troubleshooting

Fewer results than requested? Check the run summary for page, time, result or spending limits. A limit is an upper bound, not a promised row count.

No reviews returned? Some items have no public reviews, or collection may have stopped at a page, time or spending limit. Review text and star filters are not applied; inspect the coverage summary for the source outcome.

Item unavailable or blocked? Check the source URL in a browser. Restricted items and downloads are not bypassed. Other valid items in a batch can still return results.

Only search-card fields? Search returns the metadata visible on search cards. Use Get item details with the returned item identifiers for fuller public page metadata.

Memory error? Public web collection needs at least 1 GB. The default already uses 1 GB; supplied-record processing can explicitly use 256 MB.

Support

Open an issue on the Actor's Issues tab with the run ID, selected operation and a sanitised input example. This is an independent tool and is not affiliated with or endorsed by Internet Archive. Source availability and reuse rights vary by item.