We cleared 23 names and checked no domains

The app now called Quilore was called Tunelore until this week, and Miroza for about six hours before that. The renaming is not the interesting part. The interesting part is that the name was chosen by a process that was thorough about the hard problem and silent about the easy one.

What the process did well

Twenty-three candidate names went through four rounds of trademark knockout searching. It found real collisions and killed real favourites.

Anything with a recognisable audio or music root was already in commercial use, including candidates that looked clean on a casual search. Heardly is a live app on both stores. Auralis is a registered trademark for audio software. Ostinato, Auricle, Hearken, Soundlore, Harken, Reprise, Verso and Vellum are all in use by somebody. The conclusion was that the music-technology naming space is exhausted for anything that sounds like what it does, and that a pure coinage was the only reliably clear path left.

That conclusion is correct, and it is still correct today. It was also answering the wrong question.

What it did not do

The clearance document carries a list of open items. One of them reads “Live WHOIS on the chosen name across .com, .app and .fm”. It sat unticked from 19 August.

In the meantime the name went into the site constant, and from there into every canonical URL, every Open Graph tag, every sitemap entry, the share card people were meant to post, and 675 tests. The whole site was built on a domain nobody had checked.

Twenty-three names were tested for trademark collisions. None was tested for domain availability.

Both finalists were already gone

Checked properly, by RDAP against the registry:

Domain	Status	Registered
`tunelore.com`	Taken	22 March 2023, renews 2027
`miroza.com`	Taken	8 June 2011

Miroza was the corpus’s own recommended first choice, and its .com has been held since before the studio existed. Tunelore’s was registered two and a half years before we chose the name. Neither was ever available, at any point in the process that selected them.

The check that produced a wrong answer

There was one attempt, and it is worth writing down because it failed in a way that looked like success.

A curl against https://tunelore.com returned nothing, and that was read as “does not resolve, so it is probably free”. The document then recorded “domains and handles are open” and marked the claim verified.

It was wrong twice. The request failed at the TLS handshake, because a parked domain has no certificate, so the silence was about HTTPS rather than about ownership. And a domain that does not resolve is not an unregistered domain. Asked over plain HTTP, the same address answers with a bare hosting error page. It has been someone’s for three years and has simply never had a site on it.

Two lessons, and the second is the useful one:

A domain that appears dead is not a domain that is free. Registration and publication are unrelated facts.
Check the registry, not a registrar’s search box. A reseller’s “available” widget is a sales surface with an interest in the answer.

What the register looks like now

The replacement round tested 140 domains. In the first batch of thirty, twenty six were already taken.

The detail that changes how we will name the next app is not the ratio, it is the registration dates. Invented, meaningless coinages are being registered continuously: ownlore.com in June 2026, longlore.com in March 2026, verrow.com in January 2026, keeplore.com in February 2025. These are not words. Nobody is trading under them. They are inventory.

So the 2026 conclusion that coinage is the only reliably clear path left is still true for trademarks and is no longer true for domains. A fresh coinage is now roughly as likely to be blocked as a real word, and it is blocked by an owner who has no product, no mark and no intention of building one.

The binding constraint has moved from trademark clearance to domain inventory. The next naming round tests the domain first and the trademark second, which is the reverse of the order used here. Names are cheap to generate. Domains are not.

Choosing the replacement

Quilore is quill and lore, which keeps the anchor the product’s whole voice was built on: lore is accumulated knowledge, and the app is the story of one person’s listening life. It reads as bookish, which the design and the mark already were.

Two available names were rejected on their merits rather than taken because they were free. thelongaccount.com collides head on with the product’s own promise that no account is required. yearlore.com concedes the annual framing the positioning document explicitly refuses, because this is not a wrapped-style yearly summary.

Quilore has a real weakness and it is going in here rather than being argued away: the stress is ambiguous on first hearing, and it can be read KWIL-or or kwee-LOR-ay. For a product people have to type and say aloud, that is a genuine recurring cost. It was paid for a mark that is defensible.

What we are not claiming

No registry search has been run on the new name at USPTO, EUIPO or UKIPO, and no professional clearance opinion exists for any name this project has used. What exists is a knockout search and a set of direct field evidence, which is what a two-person studio can actually do, and saying otherwise would be the sort of claim this site exists to avoid.

The rename itself cost an afternoon rather than a week, for one reason: a guard test ties the address printed on the share card to the single site constant, so the two cannot drift apart without the suite going red. That test was written for a completely different reason. It is the only part of this story that went right by design.