\n In 1994, Gary Kremen registered sex dot com for $0, and in 2006 he sold\n it for over $12 million¹.\n

\n But it wasn't just NSFW domains that became gold out of thin air.\n Hundreds of early domain investors (a.k.a.{\" \"}\n domainers) registered dot coms for tens or hundreds of\n dollars—and in many cases for free—then flipped those domains for five,\n six, seven, and even eight figures.\n

\n Last week I read David Kesmodel's book about domain investing,{\" \"}\n \n The Domain Game: How People Get Rich From Internet Domain Names\n . It was published in 2008, and I bet many people who read it\n around then would have said the domain gold rush was already over. One\n book reviewer on Goodreads wrote, \"this book made clear to me that the\n domain game is pretty much over.\" From the book:\n

\n
\n Each of the 676 possible two-letter combinations and each of the\n 17,576 possible three-letter combinations was taken by March\n 2006...Each of the one thousand most common English words was\n registered.\n
\n

\n I wrote last week that{\" \"}\n \n the .com{\" \"}\n domain is already taken for all 2,315 words in the original Wordle\n answer key\n . And that includes some pretty bad domains, like wooer dot com (on\n sale for $30k if you want it). Solo domain investors bought all these\n \"good\" domains starting in the 90s, and sophisticated institutional\n investors spent hundreds of millions in the early- to mid-2000s on\n domain portfolios.\n

\n So it's too late. It's a common feeling—that you've arrived too late.\n Marc Andreessen talked about arriving to Silicon Valley in the 90s and\n feeling like he'd missed it because the PC revolution had already\n happened (...but the Web was just getting started, of course).\n

\n Ok, those two- and three-letter combos and the top thousand common\n English words were all taken by 2006...for{\" \"}\n .com. But now there are hundreds of other top-level domains\n (TLDs). You can buy hat.world{\" \"}\n for $200 right now, for example—no one has registered it yet.^{\n 2\n}{\" \"}\n So maybe it isn't too late.\n

\n The real estate metaphor is useful, I think. Investing in New York City\n real estate in the year 1990 would have already been both competitive\n and expensive. Over seven million people lived in NYC at the time, but\n less than a million lived in Shenzhen, China. Now Shenzhen has a\n population approaching 20 million, and its real estate market is\n commensurately expensive (broadly speaking—I don't know much about real\n estate). Maybe NYC is the .com here and Shenzhen is{\" \"}\n .ai{\" \"}\n (the second hottest TLD this year, according to reported secondary\n market domain sales).\n

\n Real estate isn't a perfect metaphor, of course. It starts to fray when\n comparing the finite land in the world with the near infinite set of\n possible domains. Introducing new vanity TLDs like .land\n {\" \"}\n and .xyz{\" \"}\n affects the value of existing assets, like dot coms. Would you pay $75k\n for bananabread dot com when you could buy{\" \"}\n bananabread.shop for $550 or{\" \"}\n bananabread.kitchen for $10?³\n

\n It's tricky trying to make sense of a (relatively) new asset class.\n Emerging investment opportunities often resemble predecessors, but\n there's danger in overfitting old mental models and strategies. Like,\n just look how controversial cryptocurrency has been over the last\n decade. I'm more or less a novice when it comes to investing, but it's\n way too important to the story of domains to ignore. I've never been a\n professional investor, but some of you reading this have or are. Reply!\n I'm excited to learn about this world and make sense of it all.\n

\n In my research and in the book I'll explore where domain investing is\n now as an industry and asset class. I'll cover its (sometimes shady)\n history, like registrars getting high on their own supply (i.e. buying\n domains before the customers they serve can buy them). I'll experiment\n as a domainer myself, building my own small portfolio and publicizing my\n rationale and results. I'll talk to investors, registrars, registries,\n and journalists in the domaining world. To that end, reach out if you\n have stories to share, questions you want me to pursue, or people I\n should interview (including you).\n

Footnotes

\n \n (1) Btw, no need to follow your curiosity to the browser— sex dot com\n is exactly what you think it is. I'll cover the NSFW underbelly of the\n domain world (porn) in the book. It's not something domain investors\n necessarily want to associate with or talk about, but it has been too\n lucrative to ignore.\n \n

\n \n (2) If you're wondering why $200 and not $10, it's because{\" \"}\n .world{\" \"}\n is a \"premium\" top-level domain (TLD). Domain registries (different\n from registrars) bid for TLD rights in auctions and can then charge\n what they want in turn. A company called Donuts raised over $100\n million to buy the rights to .world{\" \"}\n and many, many more TLDs, like{\" \"}\n .email. I'll cover how TLDs work in the book.\n \n

\n (3) Those are all real asking prices as of today, August 8th.\n

\n );\n}",[],[154,173,179],{"_102":187,"_95":188,"_155":189,"_157":174,"_158":136,"_159":160,"_161":162,"_104":190,"_99":32,"_164":182,"_166":191,"_167":22,"_168":-5,"_169":192,"_98":193},"52af2ecc-9ae4-11f0-80b1-0224a6c84d84","Googol.tsx","01998668-0463-7314-8932-64414b35a0f4",["D",1758896587781],"/** @jsxImportSource npm:hono@3/jsx */\n\nexport default function Googol() {\n return (\n

googol dot com

\n I am working from the Google office in NYC¹, and after my\n third trip to the cafeteria (there are four in this building) for free\n matcha, sushi, and anything else I could want to eat or drink, I decided\n now is as good a time as any to write about Google’s fingerprints on the\n secret life of domains. While writing this, I realized that it actually\n might be a better time than just any because Larry and Sergey registered\n google.com exactly 28 years ago today (I started drafting this on\n Monday): September 15, 1997.\n

\n Also today: GOOG stock hits all-time high $3 trillion.\n

\n Google is relevant to the domain world in many ways: as a (former)\n registrar, as a registry, as a way for domain investors to monetize\n parked domains, for its crackdown on{\" \"}\n domain tasting and\n {\" \"}\n kiting, and on and on.\n There’s even a good story behind the “front door of the internet,”\n Google’s apex domain google dot com.\n

Idk, Googol it

\n googol dot com is not a misspelling. google dot com\n {\" \"}\n itself is the misspelling{\" \"}\n \n according to David Koller\n {\" \"}\n who was in the room where it happened (or at least nearby).\n

\n
\n Sean and Larry were in their office, using the whiteboard, trying to\n think up a good name - something that related to the indexing of an\n immense amount of data. Sean verbally suggested the word “googolplex,”\n and Larry responded verbally with the shortened form, “googol” (both\n words refer to specific large numbers). Sean was seated at his\n computer terminal, so he executed a search of the Internet domain name\n registry database to see if the newly suggested name was still\n available for registration and use. Sean is not an infallible speller,\n and he made the mistake of searching for the name spelled as\n “google.com,” which he found to be available. Larry liked the name,\n and within hours he took the step of registering the name “google.com”\n for himself and Sergey.\n
\n

\n So Google is nerdily named after its homophone,{\" \"}\n googol, which represents ten to the power of one hundred^{\n 2\n}. Google also uses the scientific notationally named{\" \"}\n 1e100.net domain to identify its servers’ IP addresses.\n

\n In 2015, Google accidentally let google dot com{\" \"}\n expire, and ex-Googler Sanmay Ved{\" \"}\n \n purchased it\n {\" \"}\n right there on Google Domains for just $12 at 1:20am (really conjuring\n the hooded hacker image in the middle of the night). He owned it for all\n of one minute before Google reversed the transaction. Google{\" \"}\n \n paid him\n {\" \"}\n $6,006.13 for the ordeal because the number “spelled-out Google,\n numerically (squint a little and you’ll see it!).“\n

\n
\n Google Security has now contacted me, and has offered me a $x reward\n in a very Googley way. -Sanmay\n
\n

\n Sanmay is back working at Google again, and for security reasons neither\n side ever revealed how the lapse happened.\n

\n For a moment, it was possible to literally buy google dot com on google\n dot com, but nowadays it’s no longer possible to buy any{\" \"}\n domain from Google.\n

The registrar formerly known as Google Domains

\n Google comes up again and again in my interviews, and it’s most often\n about the head-scratching sale of Google Domains to Squarespace two\n years ago. One programmer friend said:\n

\n
\n I liked it when Google got into the game, because of the high trust /\n their excellent security... but then I think they gave up that\n business? Classic fucking Google\n
\n

\n Google has become notorious for{\" \"}\n killing{\" \"}\n popular products, and Google Domains was quite popular. It launched in\n 2015, and it was the{\" \"}\n \n third most popular registrar\n {\" \"}\n by the time it shut down in 2023 (10 million domains registered as of\n June 2023). I held all my domains there, and once they moved to\n Squarespace I transferred to Cloudflare.\n

\n As I’ve written about before, registrars face thin margins whereas{\" \"}\n \n registries can basically print money\n . So running a registrar is not a no-brainer for direct\n {\" \"}\n profit, but its indirect{\" \"}\n impact does seem valuable. A Hacker News commenter said that a\n registrar’s business is like the hot dogs at Costco.\n

\n For starters, any product with millions of users feels like something\n you’d hold onto. But for Google specifically, you’d think Domains would\n be valuable for upsell/cross-sell potential to adjacent products like\n Google Workspace (for a nice custom domain email). Google wants\n developers in its ecosystem, and developers love buying domains^{\n 3\n}. If you buy a domain at Google, maybe you’d also host your\n website or database there, integrate with Gemini, etc.\n

\n Many registrars do follow this strategy. My two favored registrars at\n the moment,{\" \"}\n Cloudflare\n {\" \"}\n and{\" \"}\n Vercel, both sell domains at\n cost (i.e., no markup above the wholesale registry price plus the ICANN\n fee). Vercel used to mark up domains, but now they’re at cost. A friend\n who works on domains at Vercel told me the math is clear: customers who\n buy domains there have a high lifetime value—they host on Vercel, they\n convert to Pro plans, etc. It makes sense. I create my Next.js site and\n host it on Vercel, they handle my DNS, so why not store my domains there\n too? It can also be a source of goodwill because many other registrars\n have bad UX and/or a bad reputation for sketchy behavior, like buying\n domains you’ve queried before you can.\n

\n I also think losing G Domains could be a fumble for Google in the vibe\n coding era. If an order of magnitude more people start hacking together\n apps on Gemini (or other LLM app builders like repl.it, v0.dev,\n lovable.dev, bolt.new, val.town—hey, all good non-dot-com TLDs!) then it\n would pretty handy for Google to say,{\" \"}\n do you want a domain for your cool new app?{\" \"}\n Regardless of where someone prefers to (vibe) code, registrars can meet\n people where they are (e.g., by selling domains through an API or MCP\n server). That’s not wholly a new idea—Squarespace is a website builder\n after all—but it might be a bigger opportunity now.\n

\n On the other hand, my Googler friend who I’m working with today pointed\n out that maybe Google decided the antitrust risks outweighed the\n economic and brand benefits. There are several{\" \"}\n \n Hacker News threads\n {\" \"}\n raising theories on why they sold: that it wasn’t worth the thin\n margins; that customer support was a major burden; that dealing with\n registrars and ICANN was actually a lot of work; that it was just a\n product focus thing (no wonder the Google graveyard is so big—they try\n everything); that they bartered Google Domains away to win Squarespace\n as a big Google Cloud customer and G Workspace partner; that there was\n potential friction with Google Registry; etc.\n

\n The obvious comparison is Amazon. They also have a registrar, Route 53,\n although it’s buried pretty deep in AWS (which could be a feature not a\n bug if customer support burden really is the tradeoff). It’s interesting\n to pit Amazon’s domain positioning against that of Google because, like\n Google, Amazon also has a registry, and the two often cross paths.\n

Charleston Road Registry

\n Google’s registry is incorporated as Charleston Road Registry Inc., as\n in the road that cuts through their Bay Area campus in Mountain View.\n

\n Google owns the rights to 46 top-level domains, like{\" \"}\n .dev, which they sell wholesale to registrars who in turn\n sell them to end customers. By comparison, Amazon owns 53 TLDs, and\n other registries you’ve never heard of own even more (like{\" \"}\n Identity Digital, formerly\n named Donuts and listed in the root zone database as Binky Moon, LLC).\n Registries pay Serious Money for some TLDs, like the $25 million Google\n paid for .app when it was auctioned in 2015. Google has\n {\" \"}\n applied for{\" \"}\n 101 TLDs total, so they’ve won about half of all applications/auctions.\n In alphabetical order, they own:\n

\n
\n .ads, .android, .app, .boo, .cal, .channel, .chrome, .dad, .day,\n .dclk, .dev, .docs, .drive, .eat, .esq, .fly, .foo, .gbiz, .gle,\n .gmail, .goog, .google, .guge, .hangout, .here, .how, .ing, .map,\n .meet, .meme, .mov, .new, .nexus, .page, .phd, .play, .prod, .prof,\n .rsvp, .search, .soy, .谷歌, .みんな, .グーグル, .youtube, .zip\n
\n

\n Of these, .dev{\" \"}\n is the one I see most often in the wild, and Amazon was the only other\n applicant before they withdrew. While no deals have been publicly\n disclosed, it sounds like Google and Amazon have made some backroom\n handshake deals to trade TLDs. Around the same time, Google won{\" \"}\n .drive while Amazon won .you and{\" \"}\n .talk.\n

\n About a third of these Google TLDs are tied to its products, like{\" \"}\n .gmail and{\" \"}\n .youtube. Others are kind of odd or funny, like,{\" \"}\n .dad? Google acquired .dad{\" \"}\n “to create a premiere online destination where registrants can offer and\n users can gain access to ideas, products, services and information about\n fathers and fatherhood,” but they lost out on .mom{\" \"}\n (which does exist and is owned by the{\" \"}\n \n .xyz registry\n ). Delegation of .zip triggered widespread{\" \"}\n \n security fears\n {\" \"}\n tied to potential confusion over the same-named file extension.\n

\n Each TLD (not just those owned by Google) has its own ICANN wiki page\n (e.g.,{\" \"}\n icannwiki.org/.zip), and many\n have their own regular Wikipedia page (e.g.,{\" \"}\n \n for .zip\n ). When you see a new TLD in the wild, look it up on ICANN Wiki and\n tell me about it!\n

\n As with so many domain subtopics, you could probably write an entire\n book on Google’s domain influence and involvement alone. I have to end\n these emails somewhere, though, so I’ll save Google’s role in domain\n investing (e.g., parked domains farming pennies with Google ads) among\n other topics for another day. I’ve yet to listen to part two of the{\" \"}\n Acquired{\" \"}\n podcast on Google, but I’m sure that’ll turn up even more angles to\n research. Reply with your thoughts on the lingering question:{\" \"}\n \n Why did Google really sell Google Domains, and was it a mistake?\n \n

Thanks for reading.

Footnotes

\n \n (1) Google has several offices in New York. This one is at St. John’s\n Terminal all the way west on Houston St., in the area being heavily\n rebranded as Hudson Square. I’m writing from the Greenhouse room,\n around the corner from the ROYGBIV room, and the whole building is a\n greenhouse, really. There are trees all over the north-facing\n entrance, planted on concrete cantilevers sticking out of the\n all-glass facade.\n \n

\n \n (2) When numbers get big enough they get special names, apparently.\n Googol even has alternate names like ten duotrigintillian and ten\n sexdecilliard that follow logical scale nomenclatures and, while long,\n are quite a bit shorter than writing\n 10,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000.\n \n

\n \n (3) I own a couple dozen, many of which I’ve bought for friends and\n family as birthday or holiday gifts. Ok, you say, but you’re writing a\n book about domains. Yes, that’s fair, but I was catching up with some\n former coworkers last night, and one friend said he owns ~200! I’d say\n the median number of domains owned by most dev-friends I talk to is\n about 10 (then again, I probably tend to ask more side\n project-oriented programmers than this-is-just-my-job software\n engineers).\n \n

\n );\n}",[],[154,173,188],{"_102":195,"_95":196,"_155":197,"_157":174,"_158":136,"_159":160,"_161":162,"_104":198,"_99":32,"_164":182,"_166":199,"_167":22,"_168":-5,"_169":200,"_98":201},"52af431c-9ae4-11f0-80b1-0224a6c84d84","TLDs.tsx","01998668-0463-7314-8932-69eaf8b0a573",["D",1758896587781],"/** @jsxImportSource npm:hono@3/jsx */\n\nexport default function TLDs() {\n return (\n

\n This year is the 40th anniversary of the internet’s first TLDs:{\" \"}\n .com, .org, .net,{\" \"}\n .gov, .edu, and .mil.\n

\n You know what a TLD—top-level domain—is, whether or not you\n knew the acronym. You use .com{\" \"}\n websites every day, you’ve visited country-based domains like{\" \"}\n .uk, and you might’ve seen newer ones like{\" \"}\n .xyz\n ¹. There are well over a thousand domain extensions. These\n are the “et al” in my working book title, dot com et al{\" \"}\n (which, btw, is not set in stone, so reply with your honest thoughts:\n too literary-sounding, you like the staccato, don’t like the staccato,\n too hard to say aloud, etc.).\n

So how many TLDs are there, actually?

\n I counted up 1,592 from the{\" \"}\n root zone database\n {\" \"}\n on IANA’s website. IANA is part of ICANN (mnemonic: “I can(n) give you a\n name and number on the internet”), and together they are in charge of\n domain governance. Ok cool, so that must be the true number of\n TLDs...but wait, IANA’s website also has a{\" \"}\n \n plaintext file\n {\" \"}\n updated just this morning with 1,439 TLDs. Meanwhile Domainr, the\n ICANN-accredited API provider for domains, has 1,758 TLDs on{\" \"}\n its list. It turns\n out some are retired, others are reserved for future use, and more are\n stuck in disputes².\n

Types of TLDs

Ok, so there are about 1.5k TLDs split into two main buckets:

Generic (gTLD): 1250
Country-code (ccTLD): 316

\n There are also 14 sponsored domains, 11 active test\n {\" \"}\n domains, and one infrastructure{\" \"}\n domain—.arpa—in case you’re double checking my math. And\n 169 of the set are non-Latin alphabet TLDs (called IDNs for\n Internationalized Domain Names—in Arabic, Mandarin, etc.).\n

gTLDs

\n That first batch of TLDs in 1985 brought us the household domain names\n your parents know about. Initially you had to use an appropriate\n extension for your use (.com for commercial,{\" \"}\n .net for internet infrastructure, .org{\" \"}\n for nonprofits), but registrars gave up on enforcing semantics long ago.\n

\n Over the next couple decades there was a slow trickle of new gTLDs, like\n {\" \"}\n .info and{\" \"}\n .biz. Generics got a lot more interesting in the last dozen\n or so years since ICANN opened the flood gates. We went from a couple\n dozen to over a thousand gTLDs—there’s a good chance one exists for your\n occupation, whatever it is you do (.doctor,{\" \"}\n .banker, .lawyer....dentist,{\" \"}\n .florist,{\" \"}\n .ninja). Many generics are owned by municipalities and\n companies, like .taipei and{\" \"}\n .tjmaxx, which in many cases are not open to the public for\n registration.\n

\n It costs $185k to even apply for a new gTLD plus at least $25k annually,\n and registries have bid as much as $135m in auctions for the rights to a\n single TLD³.\n

ccTLDs

\n Country-code TLDs came about in the 80s with the earliest batch of\n generics. As I’ve written about before, these are{\" \"}\n eminent domains{\" \"}\n that countries control, so when you buy one you are trusting that\n country to keep the lights on (and many haven’t, like Afghanistan,\n Syria, Libya, and Somalia).\n

\n ccTLDs are also meaningful assets in many cases, like{\" \"}\n \n Tuvalu’s .tv generating 1/12 of its gross national income\n {\" \"}\n and Anguilla’s even greater .ai{\" \"}\n windfall that has hit the mainstream—Morning Brew ran a{\" \"}\n \n feature story\n {\" \"}\n about it just this week.\n

\n As a reminder, any two-letter TLD is a ccTLD, so as you go about your\n routine keep an eye out for uses. When you see one in the wild you can\n quickly Google “.[tld] domain” to figure out what country it belongs to.\n These often turn up places I could absolutely not pick out on a map, or\n haven’t heard of even. Ask yourself, would I trust this country to\n competently and ethically operate my domain far into the future? (Or ask\n me! I like to research ccTLDs.) In most cases it’s pretty low risk.\n

Is dot com still king?

\n There’s one question I keep turning over in my head and asking others:\n how much does .com matter now that all these TLDs exist?\n

\n Does Paul Graham’s{\" \"}\n Change Your Name{\" \"}\n advice still hold? Is dot com more{\" \"}\n valuable now with a sea of diluted alternatives? There seems to be\n consensus among domain brokers I’ve spoken to or follow online that dot\n com is king, but startup founders and everyday developers like me often\n seem to prefer vanity TLDs. There could be some wishful thinking from\n both camps at play: brokers want commissions from high dollar dot com\n sales; creators want a short, cheap domain that suits their brand.\n

\n If you have an opinion or even a knee jerk reaction to the importance of\n dot com, reply to this email—I am collecting as much data as I can to\n help answer that question (and really, trying to predict the future).\n

Footnotes

\n \n (1) Cryptographic tangent: you may have also seen .eth\n {\" \"}\n addresses that look like domains. Those are part of{\" \"}\n ENS, or Ethereum Name Service, which\n is a spin on{\" \"}\n DNS\n {\" \"}\n (Domain Name System) but completely separate. Similar to website\n domain-IP pairs, ENS maps .eth{\" \"}\n domains to IP addresses for crypto wallets. Just like you visit{\" \"}\n dotcom.press{\" \"}\n instead of 216.198.79.1, you can send crypto to your drug dealer at\n {\" \"}\n trevor.eth instead of 0x0b08dA7068b73A579Bd5E8a8290f8.\n \n

\n \n (2) The root zone database has 153 “Not assigned” extensions, which\n covers the gap from IANA’s list of valid, operational TLDs\n (1592-1493=153). The unassigned include country codes like{\" \"}\n .eh{\" \"}\n for the disputed Western Sahara territory (that Morocco has tried to\n claim), branded extensions like .bananarepublic{\" \"}\n (that Banana Republic owner Gap has applied for), and test domains\n used to poke and probe DNS. Domainr’s even larger list includes some\n historically reserved testing TLDs like .example and{\" \"}\n .invalid that IANA excludes.\n \n

\n \n (3) That $135m was for .web{\" \"}\n in 2016. Was it worth it? I haven’t seen many .web-sites. $41.5m for\n {\" \"}\n .shop is the second biggest sale to date.\n \n

\n );\n}",[],[154,173,196],{"_102":203,"_95":204,"_155":205,"_157":174,"_158":136,"_159":160,"_161":162,"_104":206,"_99":32,"_164":182,"_166":207,"_167":22,"_168":-5,"_169":208,"_98":209},"52af4d3a-9ae4-11f0-80b1-0224a6c84d84","Wholesale.tsx","01998668-0463-7314-8932-6efa37bd5a4a",["D",1758896587781],"/** @jsxImportSource npm:hono@3/jsx */\n\nexport default function Wholesale() {\n return (\n

\n If you've ever bought a domain, you know what a registrar{\" \"}\n is. It's where you buy a domain. GoDaddy pays millions for 30 second\n Super Bowl ads, so actually pretty much everyone knows what a registrar\n is. What you might not know is what a registry is.\n

\n If you do know what a registry is, sorry—the point is that most people\n don't. And I'm not saying there isn't more to registrars. They do more\n than sell domains: they upsell you email hosting, website building, and\n plenty more¹. Some registrars even dabble in questionable\n behavior, like investing in their own domain portfolio using{\" \"}\n your{\" \"}\n search queries to decide which domains are sought after and therefore\n valuable. There are thousands of registrars, and their history has\n plenty of controversy and intrigue. But I'm not writing about registrars\n (for now—I'll certainly cover them later). I'm writing about their\n lesser known sibling, registries.\n

\n For every top-level domain (TLD) out there—and there are\n thousands—there's a registry that manages it. .com,{\" \"}\n .org, .xyz, .club,{\" \"}\n .ventures: they're all owned and managed by a registry. A\n company called Verisign operates both .com and{\" \"}\n .net. They sell .com domains wholesale to{\" \"}\n registrars{\" \"}\n who turn around and sell them to customers for a (thin) profit^{\n 2\n}. The current going rate Verisign charges registrars for a dot com\n is $10.26 per year.{\" \"}\n \n Verisign maintains over 170 million registrations of .com\n {\" \"}\n and .net{\" \"}\n domains, which means it earns over $1.5 billion in annual revenue from\n registrations alone.³\n {\" \"}\n (Berkshire Hathaway is a longtime Verisign investor, btw.)\n

\n I mentioned there are thousands of TLDs. .you,{\" \"}\n .music, and .now{\" \"}\n are among a dozen or so new TLDs being rolled out this year. So how do\n they come about?\n

\n Domains were something of a wild west in the early- to mid-90s, so in\n 1998 a nonprofit with a mouthful of a name was set up for domain\n governance: the Internet Corporation for Assigned Names and Numbers, or\n ICANN (pronounced \"I can\" as in \"I can help you get a name for the\n internet\"). Registries sign agreements with ICANN that grant permission\n to run a TLD so long as they meet technical, operational, and financial\n requirements.\n

\n Registries bid for rights to TLDs in auctions run by ICANN. Sometimes\n they pay a lot. Blogging platform Automattic/WordPress⁴{\" \"}\n paid $19 million for the rights to .blog{\" \"}\n in 2016. They made a big bet that bloggers would pay up for memorable\n domains. And they decide how much to charge. For example,{\" \"}\n pete.blog{\" \"}\n is a \"premium\" domain, so they're charging $500/yr for it. Whereas{\" \"}\n petemillspaugh.blog is only $20/yr. But{\" \"}\n pete.blog{\" \"}\n is available! No Pete out there wants it, especially when there are\n expressive alternatives like pete.garden for $24/yr and\n {\" \"}\n pete.football{\" \"}\n for $18/yr. Even though WordPress is still widely used, I have a hard\n time believing Automattic made back $20 million on .blog.\n

\n The range of TLD prices is kind of shocking. A registry for REITs (Real\n Estate Investment Trusts) sells .reit{\" \"}\n domains for a cool $1,000 per year. Another registry charges just $11\n yearly for .xyz{\" \"}\n domains. Back to Verisign, they've caught backlash from recent price\n hikes on{\" \"}\n .com. They have a contract allowing price increases up to\n 7% per year for four years, and they've pretty much scraped up every\n last penny of that. In their defense, secondary market sales of dot coms\n are as strong as ever, and they are a for-profit company trying to\n return money to shareholders. But it's murky. ICANN is a nonprofit whose\n mission is to \"ensure a stable, secure, and unified global Internet\",\n and ICANN is who negotiates agreements with registries like Verisign.\n But, but, but—Verisign heavily funds ICANN. Verisign is ICANN's largest\n financial backer. As anti-monopoly policy wonk Laurel Kilgour puts it:\n

\n
\n Verisign's monopoly over the .com top level domain is old enough to\n vote and drink⁵.\n
\n

\n Verisign has been around since the early internet days, and there's a\n lot more to unpack there, which I'll save for another day. In the\n meantime, what about the other registries?\n

\n For some registries, it's all they do. That registry selling{\" \"}\n .xyz{\" \"}\n domains that I mentioned earlier is called XYZ, and they own the rights\n to 35 TLDs, like .lol, .baby,{\" \"}\n .mom, .dad, .ceo, and{\" \"}\n .audio. They say on their website that they've sold over 4\n million domains. Another registry called Donuts raised over $100 million\n in venture capital to buy{\" \"}\n hundreds of TLDs\n {\" \"}\n (including what I think are some pretty good ones—like{\" \"}\n .email, .space, .band,{\" \"}\n .pizza—and ones I've used—like .life for{\" \"}\n weeksofyour.life).\n

\n But there are also registries you've actually heard of—just not for\n their registry business. Google runs a registry. So does Amazon.\n Sometimes they bid for TLDs that'll become available to the public, like\n Google paying $25 million for rights to{\" \"}\n .app, but other times they acquire TLDs for internal or\n branded use, like Amazon's .prime.\n

\n Some TLDs get stuck in closed systems and long-running litigation, like\n Amazon's ownership of{\" \"}\n .book. Amazon won the rights without any intention to sell\n {\" \"}\n .book{\" \"}\n domains to the public, which Barnes & Noble and other booksellers\n have challenged as too generic of a term to be locked away for Amazon's\n exclusive use. I would've loved to have a .book{\" \"}\n domain for this book. Google is also up against legal challenges to its\n bid for generic TLDs, like .search{\" \"}\n (which Amazon also applied for!).\n

\n I haven't even touched on registries for ccTLDs (country code top-level\n domains, like .de for Germany and .jp{\" \"}\n for Japan). Some countries run their own registries, and others partner\n with registries to capitalize on what can be a meaningful asset (as{\" \"}\n .ai has been for Anguilla, .me for Montenegro,\n {\" \"}\n .tv{\" \"}\n for Tuvalu, and many more). Anguilla switched to a partnership this year\n with a registry I mentioned earlier, Donuts (now called Identity\n Digital), to meet surging demand for its .ai{\" \"}\n domains, for example.\n

Reply

\n There's a lot about registries that I left out here in the interest of\n brevity. Reply with your unanswered questions, curiosities, and\n thoughts. Feedback helps me shape the book into something that resonates\n with you and other readers.\n

\n I'll be on a trip to Canada next week, so my next email won't be until\n September. Speaking of which, NYC friends: I'll be in town the week of\n 9/15! In the meantime, reply and forward to any friends or coworkers who\n you think would be interested in the book.\n

Footnotes

\n \n (1) WHOIS proxying, SSL certs, DNS hosting, broker services, auctions,\n etc.\n \n

\n \n (2) Or a loss! Beware the registrars who offer you promotional\n first-year pricing only to crank up the renewal costs once you're tied\n to your domain and lacking time or attention to switch.\n \n

\n \n (3) Revenue isn't quite as simple as the number of registrations\n multiplied by $10.26 because that price changes over time and domain\n buyers can lock in a price for multi-year renewals. Verisign reported\n $1.56b revenue in 2024\n (\n https://investor.verisign.com/news-releases/news-release-details/verisign-reports-fourth-quarter-and-full-year-2024-results\n ) and 170.5m registered domains at the end of Q2 2025\n (\n https://investor.verisign.com/news-releases/news-release-details/verisign-reports-second-quarter-2025-results\n ).\n \n

\n \n (4) Automattic is a WordPress hosting provider. WordPress open source\n software runs like a third to half of all websites in the world.\n Automattic CEO Matt Mullenweg has been the source of heated WordPress\n controversy of late, during which nearly 1 in 10 Automattic employees\n have resigned.{\" \"}\n \n https://theverge.com/2024/9/27/24256361/wordpress-wp-engine-drama-explained-matt-mullenweg\n .\n \n

\n \n (5) Renewing Verisign’s Contract Would Hike Domain Name Prices at 2X\n Inflation Rate, New Economic Liberties Paper Reveals. Laurel Kilgour.\n 7-29-24.{\" \"}\n \n https://www.economicliberties.us/press-release/renewing-verisigns-contract-would-hike-domain-name-prices-at-2x-inflation-rate-new-economic-liberties-paper-reveals\n .\n \n

\n );\n}",[],[154,173,204],{"_102":211,"_95":212,"_155":213,"_157":174,"_158":136,"_159":160,"_161":162,"_104":214,"_99":32,"_164":182,"_166":215,"_167":22,"_168":-5,"_169":216,"_98":217},"52af57b2-9ae4-11f0-80b1-0224a6c84d84","Wordle.tsx","01998668-0463-7314-8932-7089437aae9c",["D",1758896587781],"/** @jsxImportSource npm:hono@3/jsx */\n\nexport default function Wordle() {\n return (\n

\n Back in December, I wrote a short blog post titled{\" \"}\n dot com et al, planting a few of the seeds that are now\n growing into a book. I was writing about how all the short .com{\" \"}\n domains are already taken and wondering where that leaves us. I wrote:\n

\n
\n The main problem is availability. As an experiment, I typed four adjacent words into Cloudflare’s domain\n registrar: bean.com, beans.com, beanie.com, and{\" \"}\n beanies.com. To no surprise, none are available. But they’re not really all taken.\n {\" \"}\n beans.com serves a very simple static website.{\" \"}\n beanie.com is available for $200k, and both{\" \"}\n bean.com and beanies.com{\" \"}\n redirect. So it’s possible that all four of those are available, for a price. In my experience, the vast\n majority of dot com domains for valid English words less than about 8 characters are already spoken for.\n That’s a very unscientific hunch, but more often than not I have to think outside of the box when buying a\n domain.\n
\n

\n Now that I'm turning over domain stones full time, I decided to explore that hunch a bit more scientifically.\n To do that, I ran an experiment using the word list from Wordle¹, which felt like a goldilocks\n sample size of common 5-letter English words. For each word I wanted to know:\n

\n \n Does someone already own the .com domain?\n \n
\n If yes, does it serve a real website?\n
\n If it's for sale, what's the asking price?\n

\n There are 2,315 Wordle answers and 10,657 valid guesses in the original version of the game created by Josh\n Wardle. Josh whittled down the answer list himself, removing words most of us probably haven't heard of (like\n yabby and yobby) then randomizing the order without looking so that he and his partner could play, sans\n spoilers. Back before the New York Times bought the game, you could trivially see all past and future answers\n with a couple clicks in your browser. This wasn't an oversight by Josh, btw—you'd only spoil the game for\n yourself by looking ahead. So I found that original source code in the Wayback Machine²{\" \"}\n and copied the word list. I chose the answer list rather than the guess list because I think it's fair to say\n domains like zuppa.com, zurfs.com, and zuzim.com{\" \"}\n wouldn't be all that sought after³.\n

Taken

\n For the experiment, I first wrote a program to go through each word and check whether its .com\n {\" \"}\n domain is available. For technical reasons, I inserted a half second delay between each query⁴, so\n that means over 20 minutes of domain searching that my computer did for me while I cooked dinner (tacos).\n

\n Unsurprisingly, 0 of 2,315 Wordle solutions have an available dot com domain in the primary registrar market.\n

Parked

\n So how many of these domains serve real websites? chess.com, which{\" \"}\n \n Erik Allebest bought in 2005 for $55k in a bankruptcy sale\n , has over 8 million daily active players. awake.com redirects to Amazon because it was{\" \"}\n \n among the company names Jeff Bezos considered\n . floss.com is unused (fittingly, I guess).\n

\n To check each domain, I wrote some code to do the equivalent of entering the domain in your browser address\n bar and quickly judging if it's a real website⁵. The heuristics I used to judge what's \"real\" are\n imperfect but should be directionally correct. And even if the criteria were more precise, it'd still be hard\n to decide what's real (like the cat sticker on karma.com). Here's what I found:\n

Only a third (800ish) of the 2,315 domains serve \"real\" websites
Lots of parked domains (probably the majority) are used to solicit offers
Many sites are scraping pennies by plastering up Google ads

\n So most of these domains are unused and held by investors waiting for their big payday. How much are they\n hoping for?\n

Asking

\n Not a single Wordle domain is available for purchase directly, but we can turn to the aftermarket for the ones\n that are parked. Manually searching asking prices one by one would take too long, so I worked with a domain\n provider to search programmatically⁶. Here's what I dug up:\n

234 of the 2,315 domains (~10%) are listed for sale on the aftermarket
The average asking price is just under $600k
\n The median asking price is ~$115k (averaging atone.com for $113k and shown.com\n {\" \"}\n for ~$116.5k)\n
\n The most expensive domain is $16.7m for craze.com\n
\n The cheapest domain is just $269 for dumpy.com\n

\n Like my search for \"real\" websites, these findings are imprecise but broadly representative. I would bet a\n substantial number of domains not listed aftermarket are in fact for sale.\n

\n Other asking prices that caught my eye include lying.com for $30k, eight.com{\" \"}\n for $3.5m, and aunty.com{\" \"}\n for $60k. These prices are, of course, negotiable. I can't imagine anyone paying more for{\" \"}\n craze.com than OpenAI paid for chat.com{\" \"}\n a couple years ago, for example. In the book I'll cover how these aftermarkets work and what shady behavior\n goes on behind the scenes.\n

\n It's a bummer that so many of these pithy domains are parked by squattors running ads or waiting for a\n well-funded company to come along and make a compelling offer. I don't like it, but I don't know that it's\n unfair. It's not all that different from buying up land in the physical world and waiting around for an offer\n you can't refuse. This is something I'll keep turning over in my head as I work on the book, and I'm curious\n to hear what you think.\n

Reply

\n Thanks for reading! If you have a minute, reply and/or forward to a friend who might be interested. I'll shape\n the book around what people are interested in, so I'd love to hear what resonates, what doesn't, and what open\n questions you have. Or just reply with your Wordle score from today (I barely escaped with a 6/6).\n

Footnotes

\n \n (1) If you've never heard of or played Wordle, I'd take five and go play at{\" \"}\n nytimes.com/games/wordle. If you only have\n five minutes to either play Wordle or read my email, I'd go with Wordle. It's a wildly popular game that\n caught the world by storm, and who knows if my book will be any good.\n \n

\n \n (2) For example, you can play the Wordle puzzle from February 1, 2022 at{\" \"}\n \n web.archive.org/web/20220201/powerlanguage.co.uk/wordle\n . I won't spoil the word of that day (I guessed in 4/6), but you can spoil it for yourself by inspecting\n the source code.\n \n

\n \n (3) After writing this sentence, I naturally grew curious whether those domains were in fact sought after.\n {\" \"}\n zuppa.com and zuzim.com are owned (but unused), and zurfs.com{\" \"}\n is owned and on sale for $1,888 on GoDaddy. Go figure.\n \n

\n \n (4) For the curious programmers among you, I added the 500ms delay to play nice with Vercel's 120 req/min\n rate limit on{\" \"}\n api.vercel.com/v4/domains/status. I'm glad that Vercel exposes their domain availability search\n in a public endpoint. Few registrars do! To run more experiments like this in the future, though, I'll\n probably write a lower level bash script that uses command line tools like dig for DNS lookup.\n \n

\n \n (5) To do this, I used heuristics like: does the request return a 404 or redirect? Does it include a title\n and meta description? Is the response content sufficiently large? For future experiments, I may try feeding\n each HTML response into a classifier LLM to better gauge legitimacy.\n \n

\n \n (6) The domain search provider I worked with has not yet made their API public, but I'll plug it and share\n my code when they do.\n \n

\n );\n}",[],[154,173,212],{"_102":219,"_95":220,"_155":221,"_157":174,"_158":136,"_159":160,"_161":162,"_104":222,"_99":32,"_164":182,"_166":223,"_167":22,"_168":-5,"_169":224,"_98":225},"52af6450-9ae4-11f0-80b1-0224a6c84d84","Yu.tsx","01998668-0463-7314-8932-77adbf8b8256",["D",1758896587781],"/** @jsxImportSource npm:hono@3/jsx */\n\nexport default function Yu() {\n return (\n

Eminent domains

\n \n Eminent domain\n {\" \"}\n refers to a government’s right to seize land for public use, paying its\n owner “just compensation” in return. This power extends to a country’s\n digital land—its domain—not explicitly by law but in practice.\n

\n Every two-letter domain extension used by websites (e.g.{\" \"}\n .uk, .ca, .io,{\" \"}\n .me) belongs to a country. These extensions are called{\" \"}\n country code top-level domains, or ccTLDs, and when you buy one\n you are trusting that country to keep the lights on and allow you to\n keep using it. So really, they are eminent domains{\" \"}\n that could be taken away at any moment.\n

\n Like, just last year Afghanistan shut down many .af{\" \"}\n domains without warning, including{\" \"}\n \n queer dot af\n {\" \"}\n and Wes Bos’s bos dot af{\" \"}\n (which he talked about on a podcast episode,{\" \"}\n \n The Taliban Stole My Domain\n ). In 2021,{\" \"}\n \n Notion briefly went dark\n {\" \"}\n when its Somalian notion dot so domain stopped working.\n {\" \"}\n \n Artsy switched domains\n {\" \"}\n after its Syrian-administered art dot sy{\" \"}\n went down in 2013 during the Syrian civil war. In 2010,{\" \"}\n \n Libya started seizing .ly domains\n {\" \"}\n without notice for violating Libyan law and Islamic morality. And now\n the popular{\" \"}\n \n .io could be in jeopardy\n {\" \"}\n after Britain ceded sovereignty of the Chagos Islands to Mauritius.\n

\n Each ccTLD has its own history, none more sensational than that of{\" \"}\n .yu (that I’ve read about, at least). Yugoslavia’s{\" \"}\n .yu{\" \"}\n is dead, and its story is a dramatic crossover of geopolitical and early\n internet history.\n

Yugoslavia’s stolen, splintered, then silent .yu

\n The tl;dr is that, over the course of two decades, .yu{\" \"}\n for Yugoslavia split into .si for Slovenia,{\" \"}\n .hr for Croatia, .ba{\" \"}\n for Bosnia and Herzegovina, .mk for North Macedonia,{\" \"}\n .me for Montenegro, and .rs{\" \"}\n for Serbia. It was a long and winding road of war, theft, and\n technological innovation.\n

\n In 1989, trailblazing Slovenian computer scientist Borka Jerman-Blažič\n registered the .yu{\" \"}\n domain from her lab in Ljubljana. This was two years before the Web was\n invented, two years before the first website. Back then domains were\n used for things like email between and within academic institutions and\n research labs.\n

\n A couple years later, Jerman-Blažič (I’ll call here BJB from here on\n out) came back all excited from a computing conference in Santa Fe, NM\n and connected Yugoslavia to the global internet as soon as she could.\n That was November 1991, and it meant Yugoslavia was just the 16th\n country in the world connected to the internet. BJB said herself:\n

\n
\n Just as Vint Cerf is considered the father of the internet, one could\n say I am the mother of the internet in Yugoslavia, the mother of this\n domain.\n
\n

\n In the meantime, though, Slovenia and Croatia had declared independence\n from Yugoslavia (June of ’91). BJB reported in an interview ten years\n ago¹{\" \"}\n that Slovenia’s internet connection is how word spread of the conflict,\n leading to UN sanctions against Serbia, which in turn cut off Serbia's\n communications around Europe, leaving its people in the dark. At one\n point bombing by Yugoslav army fighter planes in Slovenia also took down\n Slovenia's email and phone lines.\n

\n After declaring independence, Slovenia got approval for{\" \"}\n .si, but before it went live they got impatient. On a\n Sunday in July ’92,{\" \"}\n \n a group of Slovenian scientists broke into BJB’s lab, stole the{\" \"}\n .yu{\" \"}\n domain records, and cut her line to the network with scissors^{\n 2\n}\n . It took BJB several months to regain access to the internet\n (which puts my experience last Friday into perspective when I was miffed\n that Spectrum was down for a few hours).\n

\n I’ll save the full story—and related ones, like the Soviet Union's\n vestigial dark corner of the web,{\" \"}\n .su—for the book, but to speedrun the rest of{\" \"}\n .yu’s splintering and ultimate silence:\n

\n 1994: Serbia (then the Federal Republic of Yugoslavia{\" \"}\n with Montenegro) gets back online by recouping .yu\n
2002: Serbia and Montenegro stop using the name Yugoslavia
2006: Montenegro declares independence
\n 2007: Serbia and Montenegro got their own ccTLDs, .rs and\n {\" \"}\n .me (which Montenegro makes good money on, btw)\n
\n 2010: .yu is officially retired\n

\n Yugoslavia came online during a time of political turmoil. The internet\n was so new that it couldn’t—and didn’t have the rules—to keep up with\n the changing of nations. The internet is a bit more resilient for it,\n but we’re left with lost artifacts:\n

\n
\n Over 4,000 websites, some of the earliest examples of internet culture\n from the region, suddenly became inaccessible via their original\n domain. For many, the deletion of .yu represented the final loss of\n the former country, the erasure of its digital identity...With the\n deletion of .yu, historians and researchers lost access to websites\n that contained important historical records. Gone are firsthand\n accounts of the NATO bombing and the Kosovo War; the mailing lists\n that scientists used to update their colleagues on the progress of the\n conflict; nostalgic forums and playful virtual nation\n experiments.³\n
\n

Footnotes

\n \n (1) BJB: “When the war began, many Slovenian scientists started\n sending emails abroad with details of what was going on here...through\n the BITNET network gateway in Paris. These emails reached various\n universities and friends with internet connections, who then informed\n the public. It allowed for a somewhat different picture of happenings\n in Yugoslavia to come out. Serbia had misfortunes, of course, due to\n its aggression against Croatia and Bosnia. This resulted in sanctions\n from the UN Security Council. Among other things, this meant cutting\n them off from the BITNET network...I warned that it would have really\n bad consequences if they disconnected the line. People in Serbia were\n already living under a constant information blockade, as the Milošević\n regime was in charge of all the TV stations. I explained how the\n computer network served as a link to the world, enabling an exchange\n of opinions and helping to inform at least the intellectual circles,\n but they wouldn’t listen to me, and the line was blocked. I believe\n this happened in late 1992.” - How does one connect a country to the\n internet? Aleksandra Domanović. 2014-05-23.{\" \"}\n \n https://rhizome.org/editorial/2014/may/23/interview-Borka-Jerman-Blazic\n .\n \n

\n \n (2) BJB: “During one weekend in July 1992, the people who were later\n employed by ARNES broke into my laboratory and, using scissors, cut\n off the line connecting my computers to the switch located in the\n cellar of the institute, then copied the zone file for .yu. The next\n day, on Monday morning, I saw what had happened, the lines being cut\n off and the computer emptied! ...After several months, I managed to\n get a new subdomain and connected my computers to the internet again.”\n Ibid.\n \n

\n \n (3) Yugoslavia’s Digital Twin. Kaloyan Kolev. 2014-05-23.{\" \"}\n \n https://www.thedial.world/articles/news/issue-9/yugolsav-wars-yu-domain-history-icann\n .\n \n

\n );\n}",[],[154,173,220],[154,173],{"_102":228,"_95":229,"_155":230,"_157":156,"_158":136,"_159":160,"_161":162,"_104":231,"_99":32,"_164":182,"_166":232,"_167":22,"_168":-5,"_169":233,"_98":234},"52af6ffe-9ae4-11f0-80b1-0224a6c84d84","Confirm.tsx","01998668-0463-7314-8932-7d0600cb6c5d",["D",1758896587781],"/** @jsxImportSource npm:hono@3/jsx */\n\ninterface ConfirmProps {\n email: string;\n token: string;\n}\n\nexport default function Confirm({ email, token }: ConfirmProps) {\n const link = `https://dotcom.press/verify?email=${email}&token=${token}`;\n\n return (\n \n \n

\n dot com et al.\n

\n The secret life of Internet domains\n

\n Click to confirm your spot on the mailing list.{\" \"}\n You can unsubscribe whenever, no hard feelings.\n

If you have a minute, reply with any domain stories or topics you'd like the book to cover.

Thanks!

-Pete

\n \n \n );\n}",[],[154,229],{"_102":236,"_95":237,"_155":238,"_157":156,"_158":136,"_159":160,"_161":162,"_104":239,"_99":32,"_164":182,"_166":240,"_167":22,"_168":-5,"_169":241,"_98":242},"52af7134-9ae4-11f0-80b1-0224a6c84d84","Newsletter.tsx","01998668-0463-7314-8932-83a88c7b5a43",["D",1758896587781],"/** @jsxImportSource npm:hono@3/jsx */\n\ninterface NewsletterProps {\n webUrl: string;\n Content: JSX.Element;\n toEmail: string;\n}\n\nexport default function Newsletter(props: NewsletterProps) {\n const { webUrl, Content, toEmail } = props;\n const unsubLink = `https://dotcom.val.run/unsubscribe?email=${toEmail}`;\n\n return (\n \n

\n \n You can also read or listen in your browser.\n \n

\n\n \n\n \n \n );\n}",[],[154,237],{"_102":244,"_95":245,"_155":246,"_157":156,"_158":136,"_159":135,"_161":162,"_104":247,"_99":248,"_164":182,"_166":249,"_167":22,"_168":-5,"_169":250,"_98":251},"3e3e9324-9ae9-11f0-b093-0224a6c84d84","Signup.tsx","01998668-0463-7314-8932-8419ba353504",["D",1758898701067],2,"/** @jsxImportSource npm:hono@3/jsx */\n\nexport default function Signup() {\n return (\n \n \n $\"Dot$ \n

dot com et al: The secret life of domains

Read email before reading the book:

\n \n \n \n \n \n );\n}",[],[154,245],[154],{"_102":254,"_95":255,"_155":256,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":257,"_99":32,"_164":165,"_166":-5,"_167":22,"_168":-5,"_169":258,"_98":283},"52af78aa-9ae4-11f0-80b1-0224a6c84d84","sqlite","01998668-0463-7314-8932-986dfe4d216b",["D",1758896587781],[259,267,275],{"_102":260,"_95":261,"_155":262,"_157":256,"_158":136,"_159":160,"_161":162,"_104":263,"_99":32,"_164":182,"_166":264,"_167":22,"_168":-5,"_169":265,"_98":266},"52af7468-9ae4-11f0-80b1-0224a6c84d84","init.ts","01998668-0463-7314-8932-8c59f73495f9",["D",1758896587781],"import { sqlite } from \"https://esm.town/v/std/sqlite\";\n\nawait sqlite.execute(`\n CREATE TABLE IF NOT EXISTS dotcom_subs (\n id INTEGER PRIMARY KEY,\n email TEXT NOT NULL UNIQUE,\n verification_token TEXT UNIQUE,\n verified_at TIMESTAMP,\n subscribed_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP\n );\n`);\n\nawait sqlite.execute(`\n CREATE TABLE IF NOT EXISTS dotcom_emails (\n id INTEGER PRIMARY KEY,\n subject TEXT NOT NULL UNIQUE,\n web_url TEXT NOT NULL UNIQUE\n );\n`);\n\nawait sqlite.execute(`\n CREATE TABLE IF NOT EXISTS dotcom_logs (\n id INTEGER PRIMARY KEY,\n subscriber_id INTEGER NOT NULL,\n newsletter_id INTEGER NOT NULL,\n sent_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,\n FOREIGN KEY (subscriber_id) REFERENCES dotcom_subs (id) ON DELETE CASCADE,\n FOREIGN KEY (newsletter_id) REFERENCES dotcom_emails (id) ON DELETE CASCADE\n );\n`);",[],[255,261],{"_102":268,"_95":269,"_155":270,"_157":256,"_158":136,"_159":160,"_161":162,"_104":271,"_99":32,"_164":182,"_166":272,"_167":22,"_168":-5,"_169":273,"_98":274},"52af7558-9ae4-11f0-80b1-0224a6c84d84","newsletters.ts","01998668-0463-7314-8932-90987a2c7d38",["D",1758896587781],"import { sqlite } from \"https://esm.town/v/std/sqlite\";\nimport Wordle from \"../jsx/newsletters/Wordle.tsx\";\nimport Gold from \"../jsx/newsletters/Gold.tsx\";\nimport Yu from \"../jsx/newsletters/Yu.tsx\";\nimport Wholesale from \"../jsx/newsletters/Wholesale.tsx\";\nimport TLDs from \"../jsx/newsletters/TLDs.tsx\";\nimport Googol from \"../jsx/newsletters/Googol.tsx\";\n\nexport const newsletters = [\n {\n jsx: Wordle,\n subject: \"Domains by Wordle\",\n webUrl: \"https://dotcom.press/archive/wordle\",\n },\n {\n jsx: Gold,\n subject: \"Domains were the new gold\",\n webUrl: \"https://dotcom.press/archive/domains-were-the-new-gold\",\n },\n {\n jsx: Yu,\n subject: \"Eminent domains\",\n webUrl: \"https://dotcom.press/archive/eminent-domains\",\n },\n {\n jsx: Wholesale,\n subject: \"Wholesale domains\",\n webUrl: \"https://dotcom.press/archive/wholesale-domains\",\n },\n {\n jsx: TLDs,\n subject: \"One thousand five hundred ninety two TLDs\",\n webUrl: \"https://dotcom.press/archive/1592-tlds\",\n },\n {\n jsx: Googol,\n subject: \"googol.com\",\n webUrl: \"https://dotcom.press/archive/googol.com\",\n },\n];\n\nfor (const { subject, webUrl } of newsletters) {\n await sqlite.execute({\n sql: `\n INSERT OR IGNORE INTO dotcom_emails (subject, web_url)\n VALUES (?, ?);\n `,\n args: [subject, webUrl],\n });\n}",[],[255,269],{"_102":276,"_95":277,"_155":278,"_157":256,"_158":136,"_159":160,"_161":162,"_104":279,"_99":32,"_164":182,"_166":280,"_167":22,"_168":-5,"_169":281,"_98":282},"52af76f2-9ae4-11f0-80b1-0224a6c84d84","read.ts","01998668-0463-7314-8932-97ba176bb189",["D",1758896587781],"import { sqlite } from \"https://esm.town/v/std/sqlite\";\n\nawait sqlite.execute({\n sql:\n `UPDATE dotcom_subs SET verified_at = CURRENT_TIMESTAMP WHERE email IN (?, ?, ?);`,\n args: [\n \"thomas.marano2@gmail.com\",\n \"lifesized@gmail.com\",\n \"umerhder@gmail.com\",\n ],\n});\n\nconsole.log(\n await sqlite.execute(`SELECT id, email, verified_at FROM dotcom_subs;`),\n);\n// console.log(await sqlite.execute(`SELECT * FROM dotcom_emails;`));\n// console.log(await sqlite.execute(`SELECT * FROM dotcom_logs;`));",[],[255,277],[255],{"_102":285,"_95":286,"_155":287,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":288,"_99":32,"_164":289,"_166":290,"_167":22,"_168":-5,"_169":291,"_98":292},"52af7a3a-9ae4-11f0-80b1-0224a6c84d84","README.md","01998668-0463-7314-8932-9fa97d7f5be8",["D",1758896587781],"file","## dotcom\n\nPowers the email signup and verification on [dotcom.press](https://dotcom.press)\n\n",[],[286],{"_102":294,"_95":295,"_155":296,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":297,"_99":32,"_164":289,"_166":298,"_167":22,"_168":-5,"_169":299,"_98":300},"52af7cba-9ae4-11f0-80b1-0224a6c84d84","logo.svg","01998668-0463-7314-8932-a3ed34da84ce",["D",1758896587781],"",[],[295],{"_102":302,"_95":303,"_155":304,"_157":-5,"_158":136,"_159":135,"_161":162,"_104":305,"_99":248,"_164":306,"_166":307,"_167":22,"_168":-5,"_169":308,"_98":309},"3e3e973e-9ae9-11f0-b093-0224a6c84d84","main.ts","01998668-0463-7314-8932-a40c96e7f8b7",["D",1758898701067],"http","import { email } from \"https://esm.town/v/std/email\";\nimport { sqlite } from \"https://esm.town/v/std/sqlite\";\nimport { Context, Hono } from \"npm:hono@3\";\nimport Confirm from \"./jsx/Confirm.tsx\";\nimport Signup from \"./jsx/Signup.tsx\";\nimport { serveFile } from \"https://esm.town/v/std/utils/index.ts\";\n\nconst app = new Hono();\n\n// Used to test signup on val town preview\napp.get(\"/\", async (c: Context) => {\n return c.html(Signup().toString());\n});\n\napp.get(\"/logo.svg\", () => serveFile(\"./logo.svg\"));\n\n// Endpoint hit by signup form on dotcom.press\napp.post(\"/signup\", async (c: Context) => {\n const formData = await c.req.formData();\n const emailAddress = formData.get(\"email\");\n if (typeof emailAddress !== \"string\") {\n return Response.json({ message: \"email is a required field\" }, {\n status: 400,\n });\n }\n\n const token = crypto.randomUUID();\n\n // Insert subscriber\n // No await: faster optimistic response on frontend\n sqlite.execute({\n sql: `\n INSERT INTO dotcom_subs (email, verification_token)\n VALUES (?, ?)\n ON CONFLICT(email)\n DO UPDATE SET\n verification_token = excluded.verification_token,\n subscribed_at = CURRENT_TIMESTAMP;\n `,\n args: [emailAddress, token],\n });\n\n // Email subscriber verification link\n // No await: faster optimistic response on frontend\n email({\n subject: \"Confirm your email for dot com et al.\",\n to: emailAddress,\n from: {\n email: \"petermillspaugh.dotcom@valtown.email\",\n name: \"Dot Com Press\",\n },\n replyTo: \"hello@dotcom.press\",\n html: (Confirm({ email: emailAddress, token }).toString()),\n });\n\n // Email myself a notification\n email({\n subject: `${emailAddress} subscribed to dot com et al.`,\n text: `${emailAddress} should verify any minute now...`,\n });\n\n return Response.json({ ok: true });\n});\n\n// Endpoint hit by verify page on dotcom.press\napp.post(\"/verify\", async (c: Context) => {\n const emailAddress = c.req.query(\"email\");\n const token = c.req.query(\"token\");\n\n // If the token is >24h old, refresh the token\n const newToken = crypto.randomUUID();\n const { rowsAffected } = await sqlite.execute({\n sql: `\n UPDATE dotcom_subs\n SET \n verification_token = ?,\n subscribed_at = CURRENT_TIMESTAMP\n WHERE email = ?\n AND verification_token = ?\n AND subscribed_at < datetime('now', '-24 hours');\n `,\n args: [newToken, emailAddress, token],\n });\n\n // Send another confirmation email if the token was refreshed\n if (rowsAffected === 1) {\n await email({\n subject: \"Confirm your email for dot com et al.\",\n to: emailAddress,\n from: {\n email: \"petermillspaugh.dotcom@valtown.email\",\n name: \"Dot Com Press\",\n },\n replyTo: \"pete@petemillspaugh.com\",\n html: (Confirm({ email: emailAddress, token: newToken }).toString()),\n });\n\n return Response.json({ refresh: true });\n }\n\n await sqlite.execute({\n sql:\n `UPDATE dotcom_subs SET verified_at = CURRENT_TIMESTAMP WHERE email = ?;`,\n args: [emailAddress],\n });\n\n // No need to await: just emailing myself a notification\n email({\n subject: `${emailAddress} verified for dot com et al.`,\n text: `${emailAddress} verified for the Dot Com Press mailing list :)`,\n });\n\n return Response.json({ ok: true });\n});\n\n// Endpoint to unsubscribe\napp.get(\"/unsubscribe\", async (c: Context) => {\n const emailAddress = c.req.query(\"email\");\n\n if (!emailAddress) {\n return c.html(`\n

Your email address is missing in the URL.

If you tried to unsubscribe and are still getting emails, send a note to hello@dotcom.press (no hard feelings).

\n `);\n }\n\n await sqlite.execute({\n sql: `\n UPDATE dotcom_subs \n SET subscribed_at = NULL \n WHERE email = ?\n `,\n args: [emailAddress],\n });\n\n // No need to await, just notifying myself\n email({\n subject: `${emailAddress} unsubscribed from dotcom.press`,\n text: `Domain nerdery isn't for everyone.`,\n });\n\n return c.html(`\n

You've successfully unsubscribed. No hard feelings.

Take care, Pete

\n `);\n});\n\nexport default app.fetch;",[],[303],{"_102":311,"_95":312,"_155":313,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":314,"_99":32,"_164":182,"_166":315,"_167":22,"_168":-5,"_169":316,"_98":317},"52af8912-9ae4-11f0-80b1-0224a6c84d84","send.ts","01998668-0463-7314-8932-aa5cbfcc78bc",["D",1758896587781],"import Newsletter from \"./jsx/Newsletter.tsx\";\nimport { newsletters } from \"./sqlite/newsletters.ts\";\nimport { email } from \"https://esm.town/v/std/email?v=11\";\nimport { sqlite } from \"https://esm.town/v/std/sqlite\";\n\nasync function sendNewsletter() {\n const newsletter = newsletters[newsletters.length - 1];\n const { subject, jsx, webUrl } = newsletter;\n\n // no-op and alert if there's already a log of the latest newsletter\n const { rows: newsletterEmailLogs } = await sqlite.execute({\n sql: `SELECT * FROM dotcom_logs WHERE newsletter_id = ?;`,\n args: [newsletters.length],\n });\n\n if (newsletterEmailLogs.length > 0) {\n await email({\n subject: `Error: already sent ${subject}`,\n html: \"Duplicate attempt to send newsletter.\",\n });\n return;\n }\n\n const { rows: subscribers } = await sqlite.execute(`\n SELECT id, email\n FROM dotcom_subs\n WHERE verified_at IS NOT NULL\n AND subscribed_at IS NOT NULL;\n `);\n\n for (const [subscriberId, emailAddress] of subscribers) {\n const { rows: subscriberEmailLogs } = await sqlite.execute({\n sql: `\n SELECT * \n FROM dotcom_logs \n WHERE newsletter_id = ?\n AND subscriber_id = ?;\n `,\n args: [newsletters.length, subscriberId],\n });\n\n // skip subscriber and alert if log exists for newsletter + subscriber\n if (subscriberEmailLogs.length > 0) {\n await email({\n subject: `Error: already sent ${subject} to subscriber`,\n html: `Duplicate send attempt for ${emailAddress}`,\n });\n continue;\n }\n\n await email({\n subject,\n to: emailAddress,\n from: {\n name: \"Dot Com Press\",\n email: \"petermillspaugh.dotcom@valtown.email\",\n },\n replyTo: \"hello@dotcom.press\",\n html: Newsletter({ Content: jsx, webUrl, toEmail: emailAddress })\n .toString(),\n });\n\n // log sent email\n await sqlite.execute({\n sql: `\n INSERT INTO dotcom_logs (newsletter_id, subscriber_id)\n VALUES (?, ?);\n `,\n args: [newsletters.length, subscriberId],\n });\n }\n}\n\nawait sendNewsletter();",[],[312],{"_102":319,"_95":320,"_155":321,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":322,"_99":32,"_164":182,"_166":323,"_167":22,"_168":-5,"_169":324,"_98":325},"52af8b56-9ae4-11f0-80b1-0224a6c84d84","test.ts","01998668-0463-7314-8932-aee460a27517",["D",1758896587781],"import { email } from \"https://esm.town/v/std/email?v=11\";\n// import { Resend } from \"npm:resend\";\nimport Newsletter from \"./jsx/Newsletter.tsx\";\nimport { newsletters } from \"./sqlite/newsletters.ts\";\n\nconst { subject, jsx, webUrl } = newsletters[newsletters.length - 1];\n\ntry {\n await email({\n subject,\n from: {\n name: \"Dot Com Press\",\n email: \"petermillspaugh.dotcom@valtown.email\",\n },\n replyTo: \"hello@dotcom.press\",\n html: Newsletter({\n webUrl,\n Content: jsx,\n toEmail: \"peterdgmillspaugh@gmail.com\",\n }).toString(),\n });\n console.log(\"test email sent\");\n} catch (e) {\n console.error(e);\n}\n\n// const resend = new Resend(Deno.env.get(\"RESEND_API_KEY\"));\n\n// const { data, error } = await resend.emails.send({\n// to: [\"peterdgmillspaugh@gmail.com\"],\n// from: \"Dot Com Press \",\n// replyTo: \"pete@dotcom.press\",\n// subject: \"Verify your email for dot com et al\",\n// text: `Hey,\n\n// Thanks for signing up to the mailing list for dot com et al: the secret life of internet domains.\n\n// To verify your email, just send a quick reply here. Past emails are archived at dotcom.press/archive.\n\n// Thanks!\n// Pete`,\n// });\n\n// error ? console.error({ error }) : console.log({ data });",[],[320],"files",[152,171,177,186,194,202,210,218,227,235,243,253,259,267,275,284,293,301,310,318],"counts",{"_330":32,"_331":160,"_332":32,"_333":334},"pulls","branches","forks","envVars",{"_335":248,"_336":248},"count","emptyCount","contents",[339,342,345,348,351,354,357],{"_102":153,"_95":154,"_155":156,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":340,"_99":32,"_164":165,"_166":-5,"_167":22,"_168":-5,"_98":341},["D",1758896587781],[154],{"_102":254,"_95":255,"_155":256,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":343,"_99":32,"_164":165,"_166":-5,"_167":22,"_168":-5,"_98":344},["D",1758896587781],[255],{"_102":285,"_95":286,"_155":287,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":346,"_99":32,"_164":289,"_166":290,"_167":22,"_168":-5,"_98":347},["D",1758896587781],[286],{"_102":294,"_95":295,"_155":296,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":349,"_99":32,"_164":289,"_166":298,"_167":22,"_168":-5,"_98":350},["D",1758896587781],[295],{"_102":302,"_95":303,"_155":304,"_157":-5,"_158":136,"_159":135,"_161":162,"_104":352,"_99":248,"_164":306,"_166":307,"_167":22,"_168":-5,"_98":353},["D",1758898701067],[303],{"_102":311,"_95":312,"_155":313,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":355,"_99":32,"_164":182,"_166":315,"_167":22,"_168":-5,"_98":356},["D",1758896587781],[312],{"_102":319,"_95":320,"_155":321,"_157":-5,"_158":136,"_159":160,"_161":162,"_104":358,"_99":32,"_164":182,"_166":323,"_167":22,"_168":-5,"_98":359},["D",1758896587781],[320],"upstreamProject",{"_102":144,"_95":362,"_110":363,"_111":364,"_108":109,"_23":365},"dotcom","https://dotcom.press | Publishing for the Internet age.","https://imagedelivery.net/iHX6Ovru0O7AjmyT5yZRoA/74cf1e3f-f8d0-4d8a-2955-f70741716b00/square512",{"_93":366,"_128":367},"petermillspaugh","https://images.clerk.dev/oauth_github/img_2TFg9SMWVOcMJQ588IzCEa1VVhI.jpeg","upstreamBranch",{"_102":139,"_104":370,"_371":372,"_95":43,"_137":144,"_99":373,"_138":-5,"_140":-5,"_374":375,"_106":-5,"_149":22,"_108":109,"_148":366,"_16":146,"_147":362},["D",1752667510190],"updatedAt",["D",1761220997043],404,"lockfileKey","1761222859609-esm.town_v_petermillspaugh_dotcom@404-main","emptyEnvVars","isNew","flags",[],"routes/_app.x.$handle.$name.environment-variables",{"_93":94,"_95":96,"_100":382,"_333":387},{"_102":103,"_95":96,"_104":383,"_106":-5,"_107":22,"_108":109,"_110":-5,"_111":-5,"_14":112,"_113":-5,"_114":22,"_115":-5,"_116":32,"_16":117,"_118":119,"_120":119,"_121":122,"_123":122,"_124":122,"_125":119,"_126":119,"_23":384,"_130":385,"_133":386,"_149":22},["D",1758896587781],{"_93":94,"_128":129},{"_93":94,"_132":129},{"_99":135,"_95":43,"_102":136,"_137":103,"_138":139,"_140":141,"_142":109,"_143":144,"_145":146,"_147":96,"_148":94},[388,394],{"_389":390,"_104":391,"_371":-5,"_110":392,"_393":162},"key","RESEND_API_KEY",["D",1758896587781],"API key for Resend","isEmpty",{"_389":395,"_104":396,"_371":-5,"_110":397,"_393":162},"VERCEL_TOKEN",["D",1758896587781],"Token for the Vercel API","actionData","errors"]

stevekrouse

petermillspaugh-dotcom

Environment variables