Ask the Archive, and watch it work

1. Open Library knows which scans of the novel are on archive.org. I'll ask for a few.

2. openlibrary_scan92 scans of the work; 5 of the first 5 checked are public

   {
     "book_id": "OL450063W",
     "limit": 5
   }

   scans of the work

   92

   • 1849 · 224 pages · public · frankensteinormo00shel_8
   • 1869 · 192 pages · public · cu31924105428902
   • 1831 · 398 pages · public · ghostseer01schiuoft
   • 1891 · 330 pages · public · frankensteinormo00shel_9
   • 1869 · 191 pages · public · frankensteinorm00shelgoog

3. The 1831 edition, ghostseer01schiuoft, is among them. Its IIIF manifest lists every scan page with its size.

4. iiif_manifest394 scan pages; the first 10 with sizes and image ids

   {
     "identifier": "ghostseer01schiuoft",
     "include_metadata": false,
     "limit": 10
   }

   title

   Frankenstein, or, The modern Prometheus

   scan pages

   394

   page 1

   1994 × 3198 px

5. Scan page 6 is the engraved frontispiece. I'll fetch it, cropped to the plate.

6. get_page_imagethe frontispiece, cropped to the plate, 503 × 820 px

   {
     "identifier": "ghostseer01schiuoft",
     "scan_page": 6,
     "region": "pct:8,13,84,84",
     "max_size": 820
   }

   
7. iiif_image_urla IIIF address for the same crop, 1,600 pixels wide

   {
     "identifier": "ghostseer01schiuoft",
     "scan_page": 6,
     "region": "pct:8,13,84,84",
     "size": "1600,"
   }

   url

   the same crop, 1,600 px wide, on iiif.archive.org

   full page

   1884 × 3074 px

8. Now the words of scan page 9, with where each one sits on the page.

9. get_hocrprinted page IX: 303 words with boxes and confidence; the first 80 kept

   {
     "identifier": "ghostseer01schiuoft",
     "scan_page": 9,
     "words_limit": 80
   }

   printed page

   IX

   words on the page

   303

   kept

   80

   • “would” · box [188, 291, 370, 340] · confidence 28
   • “speak” · box [370, 291, 540, 354] · confidence 29
   • “to” · box [540, 301, 615, 340] · confidence 30
   • … and 77 more
10. get_page_imagethe top third of printed page ix, 760 × 410 px

    {
      "identifier": "ghostseer01schiuoft",
      "scan_page": 9,
      "region": "pct:0,0,100,33",
      "max_size": 760
    }

    
11. Answer

    The 1831 edition as printed: its engraved frontispiece, and the top of page IX of the introduction with every word the OCR found.

Real calls of the server, captured 1 November 2026 (UTC), trimmed for the cards. The question and the assistant's lines are written for this demo.

1. openlibrary_editionFrankenstein, Penguin Books 2003, edition OL3701481M of work OL450063W

   {
     "edition": "9780141439471"
   }

   title

   Frankenstein

   publisher

   Penguin Books, 2003

   format

   Paperback, 273 pages

   work

   OL450063W

   scan

   frankensteinormo0000shel_d5q4

2. book_availabilityaccess: print_disabled_only; Only for people registered with the Archive's program for print-disabled readers, who borrow it for 14 days; others can't borrow it.

   {
     "book_id": "9780141439471"
   }

   access

   print_disabled_only

   other edition

   frankensteinormo00shel_8

   Only for people registered with the Archive's program for print-disabled readers, who borrow it for 14 days; others can't borrow it.

3. Not this edition, unless you're registered as a print-disabled reader. The novel is much older, though; its Open Library work counts every edition.

4. openlibrary_workfirst published 1818; 2,187 editions, 92 scans on archive.org

   {
     "work": "OL450063W"
   }

   title

   Frankenstein; or, The Modern Prometheus

   first published

   1818

   editions

   2,187

   scans on archive.org

   92

   ebook access

   public

5. Which early printings are scanned? I'll search archive.org for 1818 to 1832.

6. search_items25 archive.org texts titled Frankenstein dated 1818–1832

   {
     "query": "title:(frankenstein) AND mediatype:texts AND date:[1818 TO 1832]",
     "fields": ["identifier", "title", "date", "imagecount", "downloads"],
     "sort": ["downloads desc"],
     "limit": 10
   }

   matches

   25

   • 1818 · Frankenstein 1818 edition · 641,125 downloads
   • 1831 · Frankenstein, or, The modern Prometheus · 156,643 downloads
   • 1831 · Frankenstein; or, the Modern Prometheus - Mary Wollstonecraft Shelley · 8,381 downloads
   • 1818 · Frankenstein; or, The Modern Prometheus (1st edition, 1818, volume I) · 5,816 downloads
   • … and 21 more

7. Second in that list is the 1831 edition. Can anyone read it?

8. book_availabilityaccess: public; Public: anyone can read or download it, with no loan.

   {
     "identifier": "ghostseer01schiuoft"
   }

   access

   public

   edition

   OL7092075M

   Public: anyone can read or download it, with no loan.

9. openlibrary_cover_urla covers.openlibrary.org address (nothing downloaded)

   {
     "by": "isbn",
     "value": "9780141439471"
   }

   url

   https://covers.openlibrary.org/b/isbn/9780141439471-M.jpg

   The address only: nothing is downloaded.

10. Answer

    You can't borrow that paperback unless you're registered for print-disabled access, but the 1831 edition is free for anyone to read.

Real calls of the server, captured 1 November 2026 (UTC), trimmed for the cards. The question and the assistant's lines are written for this demo.

1. First, thirty years of captures of its front page, summarised in one call.

2. wayback_history_summary2,573 days with a capture; 1,189 answered 200, 1,376 redirected

   {
     "url": "http://geocities.yahoo.com/",
     "interval": "year",
     "one_per_day": true,
     "detail": "full",
     "limit": 5
   }

   days with a capture

   2,573

   years

   1999-2026

   answered 200

   1,189

   redirected

   1,376

3. Since 2010, 1,293 of 1,305 days with a capture were redirects, so the end is in 2009. I'll list every change of content that year.

4. wayback_cdx116 changes of content in 2009, the last 20091027225522 with status 301

   {
     "url": "http://geocities.yahoo.com/",
     "from_timestamp": "2009",
     "to_timestamp": "2010",
     "collapse": ["digest"],
     "fields": ["timestamp", "statuscode", "length"],
     "limit": 200
   }

   changes in 2009

   116

   • 26 Oct 2009, 23:31 GMT · status 200 · 5,580 bytes
   • 27 Oct 2009, 20:48 GMT · status 200 · 2,004 bytes
   • 27 Oct 2009, 22:55 GMT · status 301 · 411 bytes

5. Two captures in April sit either side of a change. Let me compare them.

6. wayback_compare+23 / −32 lines between 21 and 23 April 2009

   {
     "url": "http://geocities.yahoo.com/",
     "from_timestamp": "20090421073902",
     "to_timestamp": "20090423003922",
     "context_lines": 1
   }

   lines added

   23

   lines removed

   32

   # free web site

   Sorry, new GeoCities accounts are no longer available.
7. wayback_fetchthe page on its last day, as text

   {
     "url": "http://geocities.yahoo.com/",
     "timestamp": "20091026192158",
     "max_chars": 1500
   }

   title

   Yahoo! GeoCities: Get a web site with easy-to-use site building tools.

   GEOCITIES IS CLOSING ON OCTOBER 26, 2009.

8. wayback_fetcha day later: "Yahoo!"

   {
     "url": "http://geocities.yahoo.com/",
     "timestamp": "20091027204808",
     "max_chars": 1000
   }

   title

   Yahoo!

   Sorry, the page you requested was not found.

9. Answer

   GeoCities' front page announced the closure in April 2009 and still carried the notice on 26 October 2009; the next day the same address gave a Yahoo! “page not found”.

Real calls of the server, captured 1 November 2026 (UTC), trimmed for the cards. The question and the assistant's lines are written for this demo.

1. tvnews_program_counts345 programs on 21 channels, 51 days with at least one

   {
     "query": "artemis",
     "from_date": "2022-08-15",
     "to_date": "2022-12-15",
     "interval": "day",
     "limit": 12
   }

   programs

   345

   channels

   21

   distinct programs

   149

   days with one

   51

   • CNNW · 53 programs
   • CSPAN · 44 programs
   • KPIX · 44 programs
   • KNTV · 41 programs
   • … and 17 more

   Counts programs whose metadata matches, not mentions in captions. Air dates are UTC, from the identifiers.

2. The count is of programs whose metadata matches, not of what was said. C-SPAN's titles should say what happened when.

3. tvnews_search44 C-SPAN programs, with titles and air times

   {
     "query": "artemis",
     "channel": "CSPAN",
     "from_date": "2022-08-15",
     "to_date": "2022-12-15",
     "sort": "oldest",
     "limit": 50
   }

   C-SPAN programs

   44

   • 2022-08-29 · NASA Holds Briefing on Cancelled Artemis Launch
   • 2022-09-03 · NASA Holds Briefing After Second Cancelation of Artemis 1
   • 2022-11-16 · NASA Coverage of Artemis 1 Launch
   • 2022-11-16 · News Conference on NASA's Artemis 1 Launch
   • … and 3 more

4. “NASA Coverage of Artemis 1 Launch” aired on 16 November. Everything from that day:

5. tvnews_search31 programs on 16 November 2022, on every channel

   {
     "query": "artemis",
     "from_date": "2022-11-16",
     "to_date": "2022-11-16",
     "sort": "oldest",
     "limit": 50
   }

   programs on 16 Nov 2022

   31

   channels

   BBCNEWS, CNBC, CNNW, CSPAN, CSPAN2, FOXNEWSW, KGO, KPIX, KQED, MSNBCW

6. I'll read C-SPAN's launch coverage from five minutes in.

7. tvnews_captionscaptions from 300 s to 480 s, with a clip link

   {
     "identifier": "CSPAN_20221116_213000_NASA_Coverage_of_Artemis_1_Launch",
     "start": 300,
     "duration": 180
   }

   program

   NASA Coverage of Artemis 1 Launch

   aired

   2022-11-16 21:30:00 UTC

   caption lines

   8, 00:05:07 to 00:07:46

   [00:05:25] … 1 AND LIFTOFF. WE RISE TOGETHER, BACK TO THE MOON AND BEYOND.

8. wayback_availablethe closest capture: 20221117024858, status 200

   {
     "url": "https://www.nasa.gov/artemis-1/",
     "timestamp": "20221116"
   }

   closest capture

   20221117024858

   status

   200

9. Answer

   345 archived programs on 21 channels match “artemis” between 15 August 2022 and 15 December 2022. The launch call, as captioned on C-SPAN, is on the screen.

Real calls of the server, captured 1 November 2026 (UTC), trimmed for the cards. The question and the assistant's lines are written for this demo.

1. First, which mode the server is in. Reading needs no account.

2. account_infomode read, archive.org keys: none

   {}

   mode

   read

   archive.org keys

   none

   delete allowed

   no

3. One summary call counts the whole collection by subject, creator and year.

4. search_summary10,468 items; counts by subject, creator and year, and download figures from the 10,000 most-downloaded of 10,468 items

   {
     "query": "collection:prelinger",
     "limit": 12
   }

   items

   10,468

   top subject

   home movies (929)

   busiest years

   1948–1950 (585)

   downloads

   211,061,028+ (the 10,000 most-downloaded of 10,468 items)

5. list_collectionthe 10 most-downloaded of 10,468 items

   {
     "identifier": "prelinger",
     "fields": ["identifier", "title", "date", "downloads"],
     "limit": 10
   }

   1. About Bananas · 27,202,138
   2. Doctor in Industry (Part I) · 13,504,241
   3. Health: Your Posture · 10,637,265
   4. From the Ground Up · 10,535,911
   5. … and 6 more

6. And how many people look at it day to day.

7. item_viewsall time, 30 and 7 days, and 30 days day by day, for four identifiers

   {
     "identifiers": ["prelinger", "DuckandC1951", "AboutBan1935", "Doctorin1946"],
     "days": 30
   }

   prelinger

   210,421,082 all time · 673,019 in 30 days

   DuckandC1951

   948,570 all time · 4,349 in 30 days

   AboutBan1935

   27,202,138 all time · 1,641 in 30 days

   Doctorin1946

   13,504,241 all time · 823 in 30 days

8. Answer

   The Prelinger Archives holds 10,468 items. By year, its busiest three years are 1948–1950, with 585 items, and its most-downloaded item is “About Bananas”.

Real calls of the server, captured 1 November 2026 (UTC), trimmed for the cards. The question and the assistant's lines are written for this demo.

1. get_itemDuck and Cover, 1951; 45 files; licence field: http://creativecommons.org/licenses/publicdomain/

   {
     "identifier": "DuckandC1951",
     "files_limit": 0
   }

   title

   Duck and Cover, 1951

   made by

   Archer Productions, Inc.

   sponsor

   U.S. Federal Civil Defense Administration

   runtime

   9:15

   licence

   http://creativecommons.org/licenses/publicdomain/

2. get_subtitles2 subtitle files, SubRip (speech recognition), WebVTT (speech recognition)

   {
     "identifier": "DuckandC1951"
   }

   • DuckandC1951.asr.srt · SubRip · speech recognition
   • DuckandC1951.asr.vtt · WebVTT · speech recognition

3. Both were made by speech recognition, so expect mistakes. Reading the SubRip file:

4. get_subtitles00:00:32 to 00:08:46, 7,376 characters of machine-made text

   {
     "identifier": "DuckandC1951",
     "file": "DuckandC1951.asr.srt"
   }

   from

   00:00:32

   to

   00:08:46

   characters

   7,376

   [00:01:22] That's why these children are practicing to duck and cover just as you do in your school. We all know the atomic bomb is…

5. get_page_imagethe item's thumbnail, 180 × 124 px

   {
     "identifier": "DuckandC1951",
     "thumbnail": true
   }

   
6. get_file_links22 files: 3 videos and 19 thumbnails, with links

   {
     "identifier": "DuckandC1951",
     "formats": ["MPEG2", "Ogg Video", "h.264", "512Kb MPEG4", "Thumbnail"],
     "limit": 60
   }

   files

   22

   bytes

   335,177,627

   • MPEG2 · 256 MB
   • Ogg Video · 39 MB
   • 512Kb MPEG4 · 39 MB
   • … and 19 more
7. get_reviews81 reviews, average 4.5 stars

   {
     "identifier": "DuckandC1951",
     "limit": 1,
     "max_body_chars": 100
   }

   reviews

   81

   average

   4.5 stars

   Counts only: no review text is kept.

8. Answer

   Duck and Cover (1951), made by Archer Productions, Inc. for the U.S. Federal Civil Defense Administration. A minute of its machine transcript is on the track.

Real calls of the server, captured 1 November 2026 (UTC), trimmed for the cards. The question and the assistant's lines are written for this demo.

1. search_items1 match: Lincoln's Speech At Gettysburg (1898)

   {
     "query": "mediatype:audio AND format:SubRip AND title:(gettysburg) AND date:[1890 TO 1925]",
     "fields": ["identifier", "title", "date", "creator"],
     "limit": 5
   }

   • Lincoln's Speech At Gettysburg · Mr. W. F. Hooley · 1898-10-21
2. get_itemE. Berliner's Gramophone, 1898-10-21; 20 files; licence field: none

   {
     "identifier": "78_lincolns-speech-at-gettysburg_mr-w-f-hooley_gbia0362658a",
     "files_limit": 30
   }

   publisher

   E. Berliner's Gramophone

   date

   1898-10-21

   catalogue no.

   6012

   genre

   Recitation

   audio

   Lincoln's Speech At Gettysburg - Mr. W. F. Hooley.mp3

3. get_subtitles2 subtitle files, WebVTT (speech recognition), SubRip (copy)

   {
     "identifier": "78_lincolns-speech-at-gettysburg_mr-w-f-hooley_gbia0362658a"
   }

   • WebVTT · speech recognition
   • SubRip · copy

4. A machine-made transcript. Reading it as timed text:

5. get_subtitles00:00:01 to 00:01:47, 1,493 characters of machine-made text

   {
     "identifier": "78_lincolns-speech-at-gettysburg_mr-w-f-hooley_gbia0362658a",
     "file": "Lincoln's Speech At Gettysburg - Mr. W. F. Hooley.asr.vtt"
   }

   cues

   7, 00:00:01 to 00:01:47

   [00:00:01] [...] at Pittsburgh by Mr. William S. Dooley. Oh, sure, and seven years ago, our fathers brought forth on this continent a new nation, conceived in liberty and dedicated to the proposition that all

6. For Lincoln's text as printed, I'll search inside an 1864 book on the dedication.

7. search_inside1 match, scan page 94

   {
     "identifier": "addofhonedev00everrich",
     "query": "Fourscore"
   }

   matches

   1

   scan page

   94

   printed page

   84

   FOURSCORE and seven years ago our fathers brought forth upon this continent a new nation, conceived in Liberty, and dedicated to the proposition that all men are created equal.

8. get_full_textscan page 94, the address as printed in 1864

   {
     "identifier": "addofhonedev00everrich",
     "scan_page": 94,
     "max_chars": 2500
   }

   characters

   2,428

   paragraphs of the address

   3

9. get_page_imagethe printed page, cropped to the address, 669 × 901 px

   {
     "identifier": "addofhonedev00everrich",
     "scan_page": 94,
     "region": "pct:6,22.5,88,67",
     "max_size": 900
   }

   
10. get_page_imagethe disc's label, 640 × 640 px

    {
      "identifier": "78_lincolns-speech-at-gettysburg_mr-w-f-hooley_gbia0362658a",
      "file": "78_lincolns-speech-at-gettysburg_mr-w-f-hooley_gbia0362658a.jpg",
      "max_size": 640
    }

    
11. Answer

    Yes: Lincoln's Speech At Gettysburg, recited by Mr. W. F. Hooley for E. Berliner's Gramophone (1898). 203 of the 265 printed words appear, in order, in the machine's transcript.

Real calls of the server, captured 1 November 2026 (UTC), trimmed for the cards. The question and the assistant's lines are written for this demo.