Short answer. Buy the .com if your name is free at a normal price, because it is what people type without thinking. If it is gone, .dev is the best runner-up for anyone technical, and .me is the best one for a site built around your own name. Both renew for under 18 US dollars a year.
That is the whole decision for most people. The rest of this page is for the cases where it is not that simple: what the extensions actually cost after the first year, what developers pick and why, and whether any of it moves your rankings.
The comparison, with prices you can check
Every price below is the renewal price at Porkbun on 13 August 2026, in US dollars, because the renewal is the number you pay forever. First-year sale prices are a separate problem and get their own section.
| Extension | Renews at | Suits | The catch |
|---|---|---|---|
| .com | $11.08 | Anyone, any purpose | Your name is probably taken |
| .net | $12.52 | Fallback when .com is gone | Reads as a second choice |
| .org | $11.84 | Community projects, non-profits | Commercial sites look odd on it |
| .dev | $12.87 | Developer portfolios, side projects | HTTPS is mandatory, not optional |
| .app | $14.93 | A single app or product | Same forced HTTPS rule |
| .me | $17.27 | A site that is your name | Slightly gimmicky for a company |
| .page | $10.81 | A one-page personal site | Low recognition, and HTTPS is forced |
| .io | $51.80 | Startups, dev tools | Expensive, and a long-term question mark |
| .sh | $46.65 | Shell and CLI projects | Expensive for what it is |
| .blog | $21.11 | A writing-first site | Locks the site into being a blog |
| .xyz | $12.98 | Cheap and unrestricted | Heavily used by spam, so readers hesitate |
| .tech | $50.98 | Nothing, at that price | Costs four times .dev for less recognition |
| .codes | $57.16 | Nothing, at that price | Sold at $4.63 for year one, renews at $57.16 |
| .ai | $82.70 | AI products with funding | Two-year minimum term, so about $165 up front |
Two things jump out of that table. The gap between the cheap tier and the expensive tier is roughly five times, and it buys you nothing technical. And the extensions people reach for as a status signal, .io and .ai, are the two that cost real money every year.
The first-year price is the trap
Registrars advertise year one and quietly charge year six.
.codes is the clearest example. Porkbun sells the first year at $4.63. It renews at $57.16. That is a 12x jump, and you will not notice until the card gets charged eighteen months later.
.blog does the same thing at $2.57 for year one against $21.11 to renew. .xyz goes from $2.04 to $12.98. .online goes from $1.96 to $28.84. .io opens at $28.12 and settles at $51.80. Meanwhile .com carries no introductory offer at all, so the $11.08 you pay in year one is the $11.08 you pay next year.
When you compare domains, cover the first-year column with your thumb. The renewal price is the price.
.com does creep up, just slowly and in public. Verisign raises the wholesale fee under a contract that permits up to 7 percent in each of the last four years of a six-year cycle, and the next rise takes the wholesale price from $10.26 to $10.97 on 1 November 2026. That is an increase you can see coming years ahead, which is the opposite of the .codes situation.
What developers pick, and why .dev behaves differently
.dev has quietly become the default for personal technical sites, and it comes with a property most extensions do not have.
Google Registry put the entire .dev, .app and .page extensions on the browser HSTS preload list before anyone could register a name on them. In plain terms, HSTS preloading is a list baked into browsers that says "never speak plain HTTP to this domain." Because the whole extension is on it, HTTPS is enforced from the very first request, before your server gets a say.
Two consequences. Your site cannot be served over plain HTTP in any mainstream browser, so a misconfigured certificate means a blank error page rather than a working insecure site. And in those browsers you do not need to send the Strict-Transport-Security header yourself, because the preload list already covers you. Send it anyway: the list lives in the browser, so anything that is not a browser, like curl or a scanner, will still try plain HTTP.
For a portfolio, that is a good trade. You were going to use HTTPS anyway, and now nobody can strip it.
.io is the older prestige pick and still reads as credible in tech circles. But at $51.80 a year against $12.87, you are paying four times as much for a vibe. And there is a second reason to think twice, which is worth its own section.
The .io question nobody puts in the comparison posts
.io is not a generic extension. It is the country code for the British Indian Ocean Territory, which is the Chagos Archipelago.
On 22 May 2025 the UK signed a treaty transferring sovereignty of the Chagos Archipelago to Mauritius. ICANN published a blog post on this in November 2024 explaining the mechanism: country code extensions exist because the two-letter code sits in the ISO 3166-1 standard. If "IO" is removed from that standard, ICANN's retirement policy starts a five-year phase-out window.
Nothing has changed yet. .io is fully delegated, registrations are open, and no retirement clock is running. But the honest way to describe it is that .io has a possible expiry date and .dev does not. For a portfolio you plan to keep for a decade, that matters more than it does for a weekend project.
Does the extension affect your SEO?
For generic extensions, no. Google's own documentation on multi-regional sites is explicit: unless ICANN lists a top-level domain as a country code, Google treats any extension that resolves through the IANA root as a generic one. .dev, .xyz, .codes and .com all start from the same place.
Country codes are the exception, and this is the part that catches people out. Google uses a real country code as a geotargeting signal, which tells Search the site is aimed at that one country. Register .kw and you are telling Google you are a Kuwait site. Register .au and you are telling it you are Australian. That helps if it is true and hurts if you want international readers.
Then there is the wrinkle almost nobody knows about. Google keeps a published list of country codes it treats as generic anyway, because people use them as vanity extensions rather than geographic ones. .io, .me, .ai, .co, .cc and .tv are all on that list. So a .me portfolio is not quietly geotargeted to Montenegro, whatever a comparison post told you.
Country codes also come with paperwork. A .au domain needs a verifiable Australian presence, proven with something like an ABN, a driver's licence or an Australian trade mark, and the wholesale price rises from AUD 8.64 to AUD 9.50 excluding GST on 1 October 2026. If you are choosing a domain, you are about to need hosting to point it at, and we have written up how we pick hosting for every client project.
The ones to avoid, and why
Reader trust is the real filter here, not any technical rule.
Avoid anything with a renewal price you would not pay twice. .tech at $50.98 and .codes at $57.16 buy you nothing that .dev does not, and cost four times more.
Be careful with .xyz. It is cheap, unrestricted and technically fine, and it is exactly those properties that keep putting it in the top handful of abused extensions in Spamhaus's domain reputation reports. The registry does work on abuse, so this is a perception problem rather than a technical one. It is still a problem. A reader who has seen a hundred .xyz links in a spam folder will hesitate over yours, and hesitation is what a portfolio cannot afford.
Skip the free extensions entirely. If a registrar is giving domains away, the registry is monetising something other than your registration fee, and you have no real control over the name.
Watch out for extensions that describe your job today. A .photography domain is a problem the day you move into video. .blog has the same shape: it works until the site becomes a portfolio with a blog attached.
And check the name reads correctly when it crosses the dot. expertsexchange.com is the famous one. Say your full domain out loud before you buy it.
Something people forget: the say-it-out-loud test
The strongest argument for .com is not SEO or credibility. It is that people type it by reflex.
When someone hears your domain in a conversation, at a meetup or on a podcast, they will type the name and then .com without thinking about it. Every non-.com extension needs you to spell it out, and half your audience will still get it wrong.
That is a real cost, and it is the reason .com keeps winning despite being harder to find and slightly more expensive. It is also why a longer .com often beats a short clever alternative. abdulkadersafi.com needs no explanation. Something cut short to fit a rare extension does.
The counterargument, and it is a fair one: if your audience is developers who read your domain rather than hear it, this cost is close to zero. That is precisely the audience .dev serves.
If this is a business site, not a personal one
The calculus changes when the domain is carrying a company.
Stick to .com unless you have a strong local reason not to. Add the country code as well if you serve one market and want the geotargeting, so .com.kw alongside .com for a Kuwait business, with one canonical version and the other redirecting.
Defensive registrations start to make sense at that point, because a competitor or a phishing operator picking up the near-miss version of your name is a genuine risk rather than a hypothetical. For a personal portfolio it is money you will regret.
If you are at the stage of pricing a whole build rather than a domain, we broke down what that involves in what website development actually costs and what matters. And if the goal is to be found once the site is up, the complete guide to creating a website Google loves covers the part the domain cannot do for you.
One myth worth killing
You will read that 90 percent of domains are .com. It is not close.
Verisign publishes the numbers every quarter. In its report for the second quarter of 2026, there were 401.6 million domain registrations across all extensions, and .com stood at 166.6 million as of 30 June 2026. That is 41 percent, and it has been drifting down slowly as the newer extensions take share.
This does not weaken the case for .com. It just means the case rests on habit and recognition, which are real, rather than on a made-up statistic.
What to do now
Search for your name on .com first. If it is free and under about 15 US dollars a year, buy it and stop reading.
If it is taken, check .dev if you are technical and .me if the site is your name. Compare the renewal prices, not the first-year offers, and register for two years so you are not making this decision again in twelve months.
Then point it at hosting that will not embarrass you. That is the part that actually decides whether anyone stays on the page, and it is what our hosting service exists to handle.