AI
A 48-Hour Launch With No Playbook
On February 24, 2022, Quinn Dombrowski was on a train home, doom-scrolling news of Russia’s invasion. Two days later, a tweet from a music librarian asking for help archiving Ukrainian music collections became the trigger. Sebastian Mastovich in Vienna replied, pointing to the open-source Browsertrix web archiving crawler and a distributed approach. By March 1st, Mastovich paid out of pocket for the first servers and the project launched. No institutional approval, no roadmap, and no written guide from previous data rescue efforts existed to follow.
419 Volunteers on Day One, Zero Roadmap
By March 2nd, 419 people had joined Slack on the first day alone. The volunteer base spanned people with Slavic family connections, web archiving enthusiasts, and everyday people who saw a concrete way to help. On March 4th, the Internet Archive went down for several hours from a power outage in Northern California. That outage forced the team toward redundancy. They taught volunteers to install Docker over Zoom, walking people through the command line who had never seen a terminal. Devices ranged from old laptops to a Raspberry Pi someone set up at home.
Discovery, Triage, and the OSINT Prioritization Layer
The workflow ran: user submissions fed a giant Google Sheet, cross-referenced with Wikidata, domain intelligence, and Slavic librarians. The team initially prioritized sites hosted on Ukrainian servers, expecting Russian attacks on physical infrastructure. They quickly learned that externally hosted sites were equally at risk because refugees couldn’t always pay hosting bills. OSINT drove the crawl order after that. When missile warnings appeared for a city, volunteers virtually walked Google Maps streets looking for cultural heritage icons and checking whether those sites were already in the sheet.
18th-Century Manuscripts to After-School Dance Recitals
Archived sites ranged from 18th-century music manuscripts to the National Museum of Folk Architecture and Life, 3D virtual walkthroughs of heritage spaces, and after-school dance programs with photos from children’s recitals. Volunteers became specialists: one team focused entirely on archiving 3D walkthroughs, which required separate techniques. What existed digitally turned out to be a small fraction of Ukraine’s total cultural heritage. Ukrainian libraries kept asking for help digitizing physical objects. The team ordered scanners through Ukrainian shipping channels to avoid cross-border tariff issues, funded partly by a Pearl Jam Vitalogy Foundation grant.
8,000 Memes, a Scanner Program, and a Ukrainian-Language Gallery
The project grew beyond web archives. A gallery at gallery.sucho.org gave Ukrainian institutions a place to host digitized images using Omeka software, translated into Ukrainian for the first time. A meme wall at memes.sucho.org now holds over 8,000 war memes, each annotated with meme templates, referenced figures, and English translations. An internship placed college-age Ukrainian students into museums to build digital exhibits, aiming to cultivate the next generation of cultural heritage professionals. A web archive database at database.sucho.org links all archived content back to its metadata and provenance.
Why Communities Must Own Their Own Archives
SUCHO’s original framework: emergency web archiving is making Molotov cocktails. Quick, grassroots, outside official channels. The longer-term plan was for national libraries to act as permanent bunkers. That assumption is now strained. A parallel data rescue project from Sebastian Mastovich and Indellum has processed over 1,100 US government datasets across 86 agencies, with 150 volunteers preserving material removed from federal websites. Dombrowski’s conclusion: if you care about something, archive it yourself. A DIY web archiving zine outlines the process, from tools to mailing flash drives to friends.
Q&A
How is the data stored long-term? SUCHO migrated off AWS S3 to Wasabi, which committed to at least 5 years of storage, with institutions hosting copies in their own buckets; a plan to hand everything to UNESCO’s proposed national digital library of Ukraine stalled after a $30,000 planning process went nowhere. ▶ 27:35
Did Russian hackers try to interfere? No hate mail and no observed takedown attempts; the meme wall drew a traffic spike traced to a Russian message board post, and the biggest actual disruption was Russia attacking Ukraine’s power grid, which knocked offline sites the team was actively trying to archive. ▶ 30:42
Notable Quotes
We had anticipated the threat of, you know, Russian attacks on Ukrainian infrastructure, but we had not considered the threat of um our extremely janky Northern California power company. Quinn Dombrowski · ▶ 9:45
you’re not obligated to complete the work, but neither are you free to desist from it. Quinn Dombrowski · ▶ 20:48
Can you please please please archive it as soon as you possibly can? Quinn Dombrowski · ▶ 32:20
Key Takeaways
- SUCHO launched in 72 hours with out-of-pocket servers; speed required skipping institutional approval entirely.
- A Northern California power outage, not Russian cyberattacks, caused the first major disruption to archiving operations.
- OSINT drove crawl priority: missile alerts in a city triggered virtual Google Maps surveys for uncatalogued heritage sites.
- Externally hosted Ukrainian sites were equally at risk because refugees don’t reliably pay hosting bills.
- War memes from Telegram channels are cultural heritage; SUCHO’s 8,000-entry annotated archive is the proof.
- Governments and institutions are liabilities now; communities must archive what they care about themselves.
About the Speaker(s)
Quinn Dombrowski is a co-founder of Saving Ukrainian Cultural Heritage Online (SUCHO) and an Academic Technology Specialist in Stanford’s Division of Literatures, Cultures, and Languages and in Stanford Libraries. Their background spans medieval Slavic linguistics and computational text analysis. Quinn founded The Data-Sitters Club, a step-by-step walkthrough of computational tools and methods for literary study, and has served as co-president of the Association for Computers and the Humanities. At Stanford they teach Data Visualization with Textiles and have encoded data into physical textile work, including weaving attendance and grading records from a class and knitting a log of every regularly-scheduled meeting across 2022.