Brewster Kahle and the Internet Archive: The Library That Arrived on 44 Tapes

How a search engineer built a public memory for the web — and why a Sausalito recording studio lives inside it.

In November 1998, the Library of Congress took delivery of a piece of the internet. It arrived as 44 digital tapes: roughly two terabytes, drawn from about 500,000 websites, a snapshot called "World Wide Web 1997." The donor was Alexa Internet, a company founded two years earlier by Brewster Kahle and Bruce Gilliat.

A stack of cartridges small enough to carry held a copy of cyberspace. That contrast runs through everything Kahle built afterward.

The problem

The internet was built to move information, not to keep it. A newspaper yellows but can sit in a library for a century. A web page can vanish the day a company folds, a server shuts off, or a domain lapses. Whole communities disappear and leave nothing physical behind.

Kahle saw the problem while the public web was still young. His answer was not a better search engine. It was a library.

Before the Archive

Kahle graduated from MIT in 1982 and went to work at Thinking Machines, where the Computer History Museum places him from 1983 to 1989 as a lead engineer on the Connection Machine. In 1989 he developed WAIS, an early distributed search and retrieval system. WAIS Inc. sold to AOL in 1995. In April 1996 he founded Alexa Internet with Bruce Gilliat.

The through-line is plain. Kahle spent the 1980s on how to find information at scale. The next question was whether it would still be there to find.

Alexa crawled the public web repeatedly and, as the Library of Congress reported in 1998, donated copies of those crawls to a nonprofit Kahle had founded alongside it. Alexa did not become the Internet Archive, and not every early capture came from Alexa. But the commercial crawler and the nonprofit library grew up next to each other, and the tapes moved between them.

A library for the web

The Internet Archive was founded in 1996 and began storing web pages the same year. For five years the collection had no public reading room.

That changed on October 24, 2001, when the Wayback Machine opened at an event at the Bancroft Library in Berkeley. The launch announcement said the public could now search more than ten billion archived pages dating to 1996. In October 2025 the Archive announced the Wayback Machine had passed one trillion preserved web pages — the institution's own counting unit, and a number that keeps moving.

Kahle has kept describing the project as a library rather than a technology company. The distinction does work. A company extracts value from information. A library preserves it, organizes it, and hands it to people who were never in the room when it was made.

The television show inside the library

Computer Chronicles documented personal computing week by week, from its early years to 2002. Even the Archive's own pages disagree about when it began: a January 2026 retrospective by the Archive's Jason Scott says 1983, while the collection description says 1981. The difference likely separates early local production from the established series, but no production ledger has surfaced. I've left the start year open until one does.

On December 14, 2001 — seven weeks after the Wayback Machine launched — Kahle appeared on Net Café, a related Cheifet program. The episode survives on archive.org, and the segment runs from about 18:13 to 23:16. Producer Sarah O'Brien interviews Kahle, introduced as co-founder of the Internet Archive. He explains that the Archive had crawled the web every two months for five years onto tape, and that disk prices — four thousand dollars a terabyte — finally made public access practical. He sizes the collection at about 100 terabytes against the Library of Congress's twenty. On copyright, he says the crawler never goes past a password, and describes the collecting as "going around and just picking up all the pamphlets." Asked for a favorite find, he names the Heaven's Gate website and the White House page from September 1996.

One correction the tape itself supplies: the Archive's 2026 retrospective says Kahle demonstrated the Wayback Machine on the show. The surviving segment is an interview. He describes the Wayback Machine; nothing in the audio marks an on-screen demonstration.

By Scott's account in that 2026 retrospective, Kahle asked host Stewart Cheifet after the taping what would happen to the programs, and offered the Archive as a home. A Hewlett Foundation grant funded digitization, with Rick Prelinger involved in the work. That account is the Archive's own; the grant file behind it hasn't been checked. If it holds, the loop is a clean one: a television program recorded the arrival of the digital future, one of its guests was building a library for that future, and the library ended up preserving the program.

The readable library

Archive.org is not a catalog of titles. It is a reading room. A visitor can open a scanned book in the browser, turn its pages, and search its text.

That reading room holds millions of texts — not a catalog, but a shelf. The Texts collection alone covers books, magazines, journals, pamphlets, and periodicals contributed by libraries, publishers, and individual uploaders. A 1955 High Fidelity issue carries a Public Domain Mark and downloads freely. The Talking Machine World from 1906, contributed by the Library of Congress, downloads under a "no known restrictions" statement. A 2012 Duke University Press documentary history of early phonograph, cinema, and radio sits behind controlled lending. All three are readable. None of that makes reuse automatic — each item's rights statement governs, and public visibility is not permission.

The building

The cloud is not weightless. Digital preservation needs buildings, electricity, cooling, scanners, storage, technicians, catalogers, and replacement parts.

In April 2007 the Archive described itself as located in San Francisco's Presidio, taking in web data from Alexa and other sources while expanding into texts, audio, film, and software. In 2009, according to OpenSFHistory, it acquired a 1923 Classical Revival building at 300 Funston Avenue in the Richmond District — the former Fourth Church of Christ, Scientist, designed by Carl Werner. The Archive's own photographs from September 2009 are titled "New Building at 300 Funston."

The columned facade looks uncannily like the Archive's own temple logo. No one on the record says the resemblance drove the purchase. The building just looks the part. A hall designed for collective belief now houses collective memory, with server racks in the adjoining rooms.

More than one founder

Kahle supplied the founding idea. Others built major parts of the institution, and the Archive's own records describe their work.

Bruce Gilliat co-founded Alexa and its crawling operation, per the Library of Congress. A 2022 Archive staff profile credits Alexis Rossi, who joined in 2006, with work on the Open Library launch, four years running the Wayback Machine, and founding its webwide crawling program. The Archive identifies Mark Graham as director of the Wayback Machine, and its 2025 accounts credit his team's partnerships with repairing more than 28 million broken Wikipedia links. Jason Scott, the Archive's software curator and self-described "free-range archivist," has led the work that lets visitors run old software and Flash material in a browser. Rick Prelinger began placing his collection of ephemeral films with the Archive around 2000; by 2002, the Library of Congress reported 1,500 Prelinger titles freely viewable there.

Behind the names are digitization workers, engineers, volunteer curators, partner libraries, and individual uploaders. Some preserve millions of pages. Some save one tape that exists nowhere else.

The music library

One of the Archive's most valuable music collections did not begin inside it. Etree.org, created in 1998, was a community of tapers and collectors trading lossless concert recordings — only for artists who permitted it. By the Archive's own twentieth-anniversary account, its engineer Jon Aizen proposed in 2002 that the Archive offer that community permanent storage and bandwidth. The Live Music Archive went public in October 2002.

The Archive did not invent live-music preservation. It gave an existing preservation culture an institutional address. That story — tape trees, checksums, the Grateful Dead's taping culture, and the trip from mailed cassettes to FLAC — is its own article.

KSAN, the Record Plant, and the work that begins after preservation

Not every concert recording on archive.org belongs to the Live Music Archive. The recordings closest to the work of 2200 Studios largely sit in the Internet Archive’s radio-oriented collections.

More than fifty surviving recordings connected to the KSAN Rockbox have been located on archive.org. They document a series that ran from 1973 into the late 1970s and included Bob Marley and the Wailers, Old & In The Way, Bonnie Raitt, Fleetwood Mac, Jimmy Buffett, Journey, Peter Frampton, Tom Petty and the Heartbreakers, Warren Zevon, and many others. Listen →

The Internet Archive makes much of that audio publicly playable, but the item pages do not always provide a reliable history. Dates and personnel may come from uploaders. Rebroadcasts can be mistaken for original airings. Composite files may contain material from more than one performance. Set lists, venues, and recording details may be copied from collector notes or inherited from mislabeled tapes.

That is where the 2200 Studios research begins.

The KSAN Live at the Record Plant chronology brings together surviving audio, Raechel Donahue’s handwritten broadcast notes, station and newspaper listings, collector lineages, tape catalogs, commercial releases, spoken introductions, and alternate uploads. It checks dates and personnel against tape cards, broadcast logs, and the recordings themselves. It identifies rebroadcasts and splices, records disputed dates, distinguishes surviving recordings from broadcasts known only through documentary evidence, and reconnects the audio to the room at 2200 Bridgeway.

The division of labor is more substantial than one site storing files while another adds captions. The Internet Archive preserves and delivers many of the surviving recordings. The 2200 Studios archive reconstructs the broadcast series they came from and establishes where each recording belongs in the history of the Record Plant.

Together, they show the difference between saving an artifact and establishing its history.

Explore the complete KSAN Live at the Record Plant chronology

Ending in Sausalito

The Internet Archive does not contain everything. No library does. But it holds a trillion web pages, millions of books and recordings, a television show about the early computer age, and the sound of a band playing live in a Sausalito studio in 1973.

Somewhere in the building on Funston Avenue, a server holds the Halloween broadcast. Press play, and the room comes back on.