{ "componentChunkName": "component---src-templates-blog-post-jsx", "path": "/en/blog/general/youll-own-nothing-and-youll-be-happy-part-1/", "result": {"data":{"site":{"siteMetadata":{"siteUrl":"https://pablo.gg"}},"markdownRemark":{"id":"83fac172-3aef-52cb-8ac3-d33830cd8bb2","excerpt":"I have the same photo library in Google Photos and in a bunch of hard drives inside my apartment. This is not a workflow I recommend to normal people. This is…","html":"

I have the same photo library in Google Photos and in a bunch of hard drives inside my apartment. This is not a workflow I recommend to normal people. This is hoarder infrastructure.

\n

I do it because I don’t trust “access” the same way I trust “I have the file right here”. If you’ve ever moved countries, changed ecosystems, or watched one tech company eat another one, you know the difference matters. Convenience is nice. Ownership is nicer.

\n

And yet the whole industry spent the last decade trying to convince us those are basically the same thing. They are not. That whole “you’ll own nothing, and you’ll be happy” line was supposed to sound futuristic. Instead it became product strategy.

\n

Hoarder mode

\n

I am, objectively, the ideal customer for physical media.

\n

I like shelves. I like cartridges. I like hard drives. I like knowing that if a company wakes up on some random Tuesday and decides to kill a service, merge a platform, rename a plan, or add AI to something that used to just work, my stuff is still my stuff.

\n

This doesn’t mean I’m living in a bunker made of GameCube discs. I use Google Photos. I pay for Spotify. I stream movies. I live in the real world. But I don’t confuse convenience with control.

\n

So yes, all my photos are in Google Photos. And yes, all my photos are also stored locally. Because when a company says “don’t worry, it’s in the cloud”, what I hear is “future Pablo, this is your problem now.” (Also: future Pablo, link the degoogling post here when you finally write it.)

\n

Concord was the cleanest version of the problem

\n

Concord was such a disaster it almost looped back into being interesting.

\n

Sony launched it, it bombed, refunded everyone, and then removed the digital version from players’ accounts. Depending on how you experienced it, the install either disappeared from the console or turned into a dead icon you couldn’t do anything with. Same result. A thing people had “bought” stopped existing the moment Sony decided the experiment was over.

\n

Yes, people got the money back. No, that is not the point.

\n

The point is that Sony showed, with zero subtlety, what a digital purchase actually is. Not ownership. Permission. Temporary, revocable, account-bound permission.

\n

Now compare that with P.T.. Konami pulled it from the PlayStation Store after the Kojima breakup, which already sucked, but if you had downloaded it before that, it stayed on your console. People held on to those PS4s like they were carrying cursed treasure. That was bad enough already. Concord managed to make it worse.

\n

The old version of digital was “you can’t buy it anymore”. The new version is “you can’t even keep the useless dead file you already had”. Amazing progress.

\n

Anime people already learned this lesson

\n

Gamers are not even the first victims here. Anime fans got a speedrun version of the same thing when Funimation got folded into Crunchyroll.

\n

People had digital copies tied to Funimation that simply did not make the trip. Not “please wait while we migrate your library”. Not “we’re sorting out licensing details”. Just gone. Crunchyroll’s own support page says those digital copies are not available there anymore.

\n

Back in the DVD days, this problem did not exist. You bought the disc, threw away the plastic wrap, put it on a shelf, and that was the end of the negotiation. The publisher could disappear, get acquired, or explode in a corporate merger and your movie would still be sitting there, ready to go.

\n

That is what ownership looks like. Boring. Reliable. Beautiful.

\n

The subscription math is fake-good

\n

Game Pass and PS Plus are the kind of deal that sounds amazing until you stop looking at the monthly price and start looking at the years.

\n

Let’s say you pay around $200 a year for a premium game subscription. In ten years, that is $2,000. And what do you have at the end of it? A receipt history and some fond memories, I guess.

\n

You do not have a shelf of games. You do not have something you can resell. You do not have the specific 2026 version of that game before three balance patches and two monetization experiments turned it into something else. The second you stop paying, the whole library evaporates.

\n

People hear this and say “yeah but I don’t replay games”. Fair. Until the service gets worse.

\n

Because that is the real checkmate here. The trap is not that you personally need to own every movie or every game forever. The trap is that once enough people stop owning things, companies get to make the service worse and you have nowhere good to go. More ads, higher prices, smaller catalog, worse support, more lock-in. The usual enshittification arc.

\n

If you want a stupidly specific example, look at the 3DS. Pokemon Shuffle was a real Nintendo game on a real Nintendo handheld and now, legally, it may as well be smoke for anyone who didn’t grab it in time. The Legend of Zelda: Four Swords Anniversary Edition was free and still managed to become unobtainable. Free. Gone. Nintendo somehow found a way to make zero euros expire.

\n

Nintendo is now importing digital nonsense into physical media

\n

Nintendo used to be the company of “blow on the cartridge and try again.” Very physical. Very stupid. Very honest.

\n

Now we get stuff like Super Mario 3D All-Stars being sold with artificial scarcity so the second-hand price climbs to around €130 for no good reason beyond Nintendo deciding FOMO should have a retail box. I already wrote about how this kind of pricing logic works in the collectibles world here. It is annoying there too, but at least nobody pretends it is pro-consumer.

\n

Then came the Switch 2 game-key cards. Nintendo’s own support page literally says the card does not contain the full game data. It is the key that lets you download the actual game.

\n

So now even when you buy the thing in a store, take the box home, and insert the card in the console, you may still be buying a permission slip. Physical-looking DRM. A cartridge cosplay.

\n

Nintendo says you only need the internet the first time, which is better than always-online garbage, sure. But the preservation problem is still right there. If the real game lives on a server first, that physical product has an expiration date hiding inside it.

\n

Cartridges used to be the whole point. Now sometimes they’re just a receipt with extra steps.

\n

Your phone is also turning into a frontend

\n

I miss when phones had features that actually lived inside the phone.

\n

My old Nokia had Wi-Fi and could do some smart things, but most of the time it was just a solid little brick that worked offline and minded its own business. Now every launch event is full of magical features that depend on some cloud service quietly doing the real work somewhere else.

\n

That means the feature is not really part of the device. It is part of a service attached to the device.

\n

The day the company kills that service, moves it behind a subscription tier, or decides only the new model deserves it, your very fancy hardware suddenly forgets how to do the thing from the commercial. Congrats on the purchase. You bought a screen for a server.

\n

I am not a purist, just tired

\n

I am not telling you to cancel every subscription, throw your phone in a canal, and live off ripped DVDs and local MP3s like it’s 2007. I still make compromises. Everybody does.

\n

But I do think a few rules still make sense:

\n\n

Also, physical media is not just better for preservation. It is often better for your wallet. You can buy a used game, finish it, and sell it to the next guy. Try doing that with a PSN license.

\n

Part 2 is going deeper into software subscriptions specifically, because that rabbit hole gets even dumber. Fitness apps, Adobe, “AI features”, apps charging monthly fees to write text into SQLite, all the usual modern nonsense. I already touched that nerve a bit in my Musclog post about nutrition tracking and scammy fitness app subscriptions, but I have more to say.

\n

For now, that’s the whole rant: cloud stuff is convenient, subscriptions are sometimes unavoidable, and ownership still matters. Especially when the companies telling you it doesn’t are the exact ones with a financial incentive to make sure you never really keep anything again.

\n

See you in part 2.

","fields":{"postHashId":"Z2VuZXJhbHRydWVudWxsMjAyNi0wNS0xNlQwMDowMDowMC4wMDBa","slug":"/2026/2026-05-16-youll-own-nothing-and-youll-be-happy-part-1.en/","path":"/blog/general/youll-own-nothing-and-youll-be-happy-part-1/","locale":"en"},"readingTime":{"minutes":7.41},"frontmatter":{"path":"youll-own-nothing-and-youll-be-happy-part-1","allowComments":true,"title":"You'll own nothing and you'll be happy - Part 1","date":"2026-05-16T00:00:00.000Z","categories":["general"],"tags":["subscriptions","digital-ownership","games","physical-media","enshittification"],"hideExcerpt":false,"subtitle":"Cloud convenience is cute until somebody decides your purchase no longer exists"}},"categoryImage":{"childImageSharp":{"original":{"width":1920,"height":1080,"src":"/static/categories_general-2e01bc51fc29da41beb68a86d5ec193f.jpg"}}}},"pageContext":{"postHashId":"Z2VuZXJhbHRydWVudWxsMjAyNi0wNS0xNlQwMDowMDowMC4wMDBa","relatedPosts":[{"fields":{"postHashId":"Z2FtZXN0cnVlbnVsbDIwMjQtMTAtMjFUMDA6MDA6MDAuMDAwWg==","slug":"/2024/2024-10-21-a-guide-to-retro-game-shopping-in-seoul.en/","path":"/blog/games/a-guide-to-retro-game-shopping-in-seoul/","locale":"en"},"frontmatter":{"tags":["games","shopping","seoul","korea","action figures","retro games","video games"],"categories":["games"],"allowComments":true,"publishOnMedium":false,"cover":null,"date":"2024-10-21T00:00:00.000Z","id":null,"path":"a-guide-to-retro-game-shopping-in-seoul","show":true,"title":"A guide to retro game shopping in Seoul","hideExcerpt":false,"subtitle":"A guide to the best places to buy retro games and action figures in Seoul."}},{"fields":{"postHashId":"Z2FtZXN0cnVlbnVsbDIwMjItMDktMjdUMDA6MDA6MDAuMDAwWg==","slug":"/2022/2022-09-27-e3-2023-dates-revealed.en/","path":"/blog/games/e3-2023-dates-revealed/","locale":"en"},"frontmatter":{"tags":["e3","los angeles","games","nintendo","sony","microsoft"],"categories":["games"],"allowComments":true,"publishOnMedium":false,"cover":null,"date":"2022-09-27T00:00:00.000Z","id":null,"path":"e3-2023-dates-revealed","show":true,"title":"E3 2023 dates revealed","hideExcerpt":false,"subtitle":"LET'S GOOOOOOOO"}},{"fields":{"postHashId":"Z2FtZXN0cnVlbnVsbDIwMjItMDYtMDFUMDA6MDA6MDAuMDAwWg==","slug":"/2022/2022-06-01-10-years-anniversary-of-my-first-e3.en/","path":"/blog/games/10-years-anniversary-of-my-first-e3/","locale":"en"},"frontmatter":{"tags":["e3","e3 2012","games","los angeles","lacc","convention center","nintendo","sony","microsoft","gameblast","media","journalism"],"categories":["games"],"allowComments":true,"publishOnMedium":false,"cover":null,"date":"2022-06-01T00:00:00.000Z","id":null,"path":"10-years-anniversary-of-my-first-e3","show":true,"title":"10 years anniversary of my first E3","hideExcerpt":false,"subtitle":"10 years ago today I was embarking on a journey to my first E3 conference."}},{"fields":{"postHashId":"Z2FtZXN0cnVlbnVsbDIwMjEtMTEtMThUMDA6MDA6MDAuMDAwWg==","slug":"/2021/2021-11-18-how-to-load-assets-asynchronously-with-phaser-3.en/","path":"/blog/games/how-to-load-assets-asynchronously-with-phaser-3/","locale":"en"},"frontmatter":{"tags":["coding","games","javascript","phaser","phaser 3","asynchronous","rexrainbow","phaser loader"],"categories":["games"],"allowComments":true,"publishOnMedium":false,"cover":null,"date":"2021-11-18T00:00:00.000Z","id":null,"path":"how-to-load-assets-asynchronously-with-phaser-3","show":true,"title":"How to load assets asynchronously with Phaser 3","hideExcerpt":false,"subtitle":"Asynchronous asset loading in Phaser 3's preload function."}},{"fields":{"postHashId":"Y29kaW5ndHJ1ZW51bGwyMDIxLTEwLTEyVDAwOjAwOjAwLjAwMFo=","slug":"/2021/2021-10-12-i-made-a-top-down-game-version-of-my-blog-with-phaser-and-react.en/","path":"/blog/coding/i-made-a-top-down-game-version-of-my-blog-with-phaser-and-react/","locale":"en"},"frontmatter":{"tags":["coding","games","javascript","phaser","phaser 3","react","blog","rpg","rpg maker","top-down"],"categories":["coding"],"allowComments":true,"publishOnMedium":null,"cover":null,"date":"2021-10-12T00:00:00.000Z","id":null,"path":"i-made-a-top-down-game-version-of-my-blog-with-phaser-and-react","show":true,"title":"I made a top-down game version of my blog with Phaser and React","hideExcerpt":null,"subtitle":"Learn how to create a top-down RPG game using Phaser and React, including integrating with Gatsby and creating the game map with Tiled"}},{"fields":{"postHashId":"Y29kaW5ndHJ1ZW51bGwyMDIxLTEwLTA4VDAwOjAwOjAwLjAwMFo=","slug":"/2021/2021-10-08-how-to-create-a-top-down-rpg-maker-like-game-with-phaser-js-and-react.en/","path":"/blog/coding/how-to-create-a-top-down-rpg-maker-like-game-with-phaser-js-and-react/","locale":"en"},"frontmatter":{"tags":["coding","games","javascript","phaser","phaser 3","react","dialog box","rpg","rpg maker"],"categories":["coding"],"allowComments":true,"publishOnMedium":null,"cover":null,"date":"2021-10-08T00:00:00.000Z","id":null,"path":"how-to-create-a-top-down-rpg-maker-like-game-with-phaser-js-and-react","show":true,"title":"How to create a top-down RPG Maker like game with Phaser JS and React","hideExcerpt":null,"subtitle":"From RPG Maker dreams to Phaser reality: a game dev journey."}},{"fields":{"postHashId":"Y29kaW5ndHJ1ZW51bGwyMDIxLTEwLTA2VDAwOjAwOjAwLjAwMFo=","slug":"/2021/2021-10-06-creating-a-dialog-box-with-react-for-a-phaser-game.en/","path":"/blog/coding/creating-a-dialog-box-with-react-for-a-phaser-game/","locale":"en"},"frontmatter":{"tags":["coding","games","javascript","phaser","phaser 3","react","dialog box","rpg","rpg maker","material ui"],"categories":["coding"],"allowComments":true,"publishOnMedium":null,"cover":null,"date":"2021-10-06T00:00:00.000Z","id":null,"path":"creating-a-dialog-box-with-react-for-a-phaser-game","show":true,"title":"Creating a dialog box with React for a Phaser game","hideExcerpt":null,"subtitle":"Phaser game UI: Blend canvas and DOM."}},{"fields":{"postHashId":"Z2FtZXN0cnVlbnVsbDIwMjEtMTAtMDJUMDA6MDA6MDAuMDAwWg==","slug":"/2021/2021-10-02-automatically-update-the-embedded-tileset-for-all-your-tiled-game-with-a-nodejs-script.en/","path":"/blog/games/automatically-update-the-embedded-tileset-for-all-your-tiled-game-with-a-nodejs-script/","locale":"en"},"frontmatter":{"tags":["coding","games","javascript","phaser","phaser 3","nodejs","node","tiled","map","tilemap","tileset"],"categories":["games"],"allowComments":true,"publishOnMedium":null,"cover":null,"date":"2021-10-02T00:00:00.000Z","id":null,"path":"automatically-update-the-embedded-tileset-for-all-your-tiled-game-with-a-nodejs-script","show":true,"title":"Automatically update the embedded Tileset for all your Tiled maps a Node.js script","hideExcerpt":null,"subtitle":"Load Tiled maps in Phaser JS easily."}},{"fields":{"postHashId":"Z2FtZXN0cnVlbnVsbDIwMjEtMDktMDlUMDA6MDA6MDAuMDAwWg==","slug":"/2021/2021-09-09-creating-a-mobile-build-of-my-phaser-js-game-game-devlog-23.en/","path":"/blog/games/creating-a-mobile-build-of-my-phaser-js-game-game-devlog-23/","locale":"en"},"frontmatter":{"tags":["coding","games","javascript","phaser","phaser 3","game devlog","gamedev","skate platformer","super ollie vs pebble corp","webpack","tiled","mobile","android","mobile build"],"categories":["games"],"allowComments":true,"publishOnMedium":false,"cover":null,"date":"2021-09-09T00:00:00.000Z","id":null,"path":"creating-a-mobile-build-of-my-phaser-js-game-game-devlog-23","show":true,"title":"Creating a mobile build of my Phaser JS game - Game Devlog #23","hideExcerpt":false,"subtitle":"Porting Super Ollie to mobile with virtual controls."}},{"fields":{"postHashId":"Z2FtZXN0cnVlbnVsbDIwMjEtMDgtMjhUMDA6MDA6MDAuMDAwWg==","slug":"/2021/2021-08-28-creating-moving-platforms-for-my-phaser-js-game-game-devlog-22.en/","path":"/blog/games/creating-moving-platforms-for-my-phaser-js-game-game-devlog-22/","locale":"en"},"frontmatter":{"tags":["coding","games","javascript","phaser","phaser 3","game devlog","gamedev","skate platformer","super ollie vs pebble corp","webpack","tiled","moving platforms","platforms"],"categories":["games"],"allowComments":true,"publishOnMedium":false,"cover":null,"date":"2021-08-28T00:00:00.000Z","id":null,"path":"creating-moving-platforms-for-my-phaser-js-game-game-devlog-22","show":true,"title":"Creating moving platforms for my Phaser JS game - Game Devlog #22","hideExcerpt":false,"subtitle":"Super Ollie: Making moving platforms, finally."}}],"alternativeHtml":"

I have the same photo library in Google Photos and in a bunch of hard drives inside my apartment. This is not a workflow I recommend to normal people. This is hoarder infrastructure.

\n

I do it because I don’t trust “access” the same way I trust “I have the file right here”. If you’ve ever moved countries, changed ecosystems, or watched one tech company eat another one, you know the difference matters. Convenience is nice. Ownership is nicer.

\n

And yet the whole industry spent the last decade trying to convince us those are basically the same thing. They are not. That whole “you’ll own nothing, and you’ll be happy” line was supposed to sound futuristic. Instead it became product strategy.

\n

Hoarder mode

\n

I am, objectively, the ideal customer for physical media.

\n

I like shelves. I like cartridges. I like hard drives. I like knowing that if a company wakes up on some random Tuesday and decides to kill a service, merge a platform, rename a plan, or add AI to something that used to just work, my stuff is still my stuff.

\n

This doesn’t mean I’m living in a bunker made of GameCube discs. I use Google Photos. I pay for Spotify. I stream movies. I live in the real world. But I don’t confuse convenience with control.

\n

So yes, all my photos are in Google Photos. And yes, all my photos are also stored locally. Because when a company says “don’t worry, it’s in the cloud”, what I hear is “future Pablo, this is your problem now.” (Also: future Pablo, link the degoogling post here when you finally write it.)

\n

Concord was the cleanest version of the problem

\n

Concord was such a disaster it almost looped back into being interesting.

\n

Sony launched it, it bombed, refunded everyone, and then removed the digital version from players’ accounts. Depending on how you experienced it, the install either disappeared from the console or turned into a dead icon you couldn’t do anything with. Same result. A thing people had “bought” stopped existing the moment Sony decided the experiment was over.

\n

Yes, people got the money back. No, that is not the point.

\n

The point is that Sony showed, with zero subtlety, what a digital purchase actually is. Not ownership. Permission. Temporary, revocable, account-bound permission.

\n

Now compare that with P.T.. Konami pulled it from the PlayStation Store after the Kojima breakup, which already sucked, but if you had downloaded it before that, it stayed on your console. People held on to those PS4s like they were carrying cursed treasure. That was bad enough already. Concord managed to make it worse.

\n

The old version of digital was “you can’t buy it anymore”. The new version is “you can’t even keep the useless dead file you already had”. Amazing progress.

\n

Anime people already learned this lesson

\n

Gamers are not even the first victims here. Anime fans got a speedrun version of the same thing when Funimation got folded into Crunchyroll.

\n

People had digital copies tied to Funimation that simply did not make the trip. Not “please wait while we migrate your library”. Not “we’re sorting out licensing details”. Just gone. Crunchyroll’s own support page says those digital copies are not available there anymore.

\n

Back in the DVD days, this problem did not exist. You bought the disc, threw away the plastic wrap, put it on a shelf, and that was the end of the negotiation. The publisher could disappear, get acquired, or explode in a corporate merger and your movie would still be sitting there, ready to go.

\n

That is what ownership looks like. Boring. Reliable. Beautiful.

\n

The subscription math is fake-good

\n

Game Pass and PS Plus are the kind of deal that sounds amazing until you stop looking at the monthly price and start looking at the years.

\n

Let’s say you pay around $200 a year for a premium game subscription. In ten years, that is $2,000. And what do you have at the end of it? A receipt history and some fond memories, I guess.

\n

You do not have a shelf of games. You do not have something you can resell. You do not have the specific 2026 version of that game before three balance patches and two monetization experiments turned it into something else. The second you stop paying, the whole library evaporates.

\n

People hear this and say “yeah but I don’t replay games”. Fair. Until the service gets worse.

\n

Because that is the real checkmate here. The trap is not that you personally need to own every movie or every game forever. The trap is that once enough people stop owning things, companies get to make the service worse and you have nowhere good to go. More ads, higher prices, smaller catalog, worse support, more lock-in. The usual enshittification arc.

\n

If you want a stupidly specific example, look at the 3DS. Pokemon Shuffle was a real Nintendo game on a real Nintendo handheld and now, legally, it may as well be smoke for anyone who didn’t grab it in time. The Legend of Zelda: Four Swords Anniversary Edition was free and still managed to become unobtainable. Free. Gone. Nintendo somehow found a way to make zero euros expire.

\n

Nintendo is now importing digital nonsense into physical media

\n

Nintendo used to be the company of “blow on the cartridge and try again.” Very physical. Very stupid. Very honest.

\n

Now we get stuff like Super Mario 3D All-Stars being sold with artificial scarcity so the second-hand price climbs to around €130 for no good reason beyond Nintendo deciding FOMO should have a retail box. I already wrote about how this kind of pricing logic works in the collectibles world here. It is annoying there too, but at least nobody pretends it is pro-consumer.

\n

Then came the Switch 2 game-key cards. Nintendo’s own support page literally says the card does not contain the full game data. It is the key that lets you download the actual game.

\n

So now even when you buy the thing in a store, take the box home, and insert the card in the console, you may still be buying a permission slip. Physical-looking DRM. A cartridge cosplay.

\n

Nintendo says you only need the internet the first time, which is better than always-online garbage, sure. But the preservation problem is still right there. If the real game lives on a server first, that physical product has an expiration date hiding inside it.

\n

Cartridges used to be the whole point. Now sometimes they’re just a receipt with extra steps.

\n

Your phone is also turning into a frontend

\n

I miss when phones had features that actually lived inside the phone.

\n

My old Nokia had Wi-Fi and could do some smart things, but most of the time it was just a solid little brick that worked offline and minded its own business. Now every launch event is full of magical features that depend on some cloud service quietly doing the real work somewhere else.

\n

That means the feature is not really part of the device. It is part of a service attached to the device.

\n

The day the company kills that service, moves it behind a subscription tier, or decides only the new model deserves it, your very fancy hardware suddenly forgets how to do the thing from the commercial. Congrats on the purchase. You bought a screen for a server.

\n

I am not a purist, just tired

\n

I am not telling you to cancel every subscription, throw your phone in a canal, and live off ripped DVDs and local MP3s like it’s 2007. I still make compromises. Everybody does.

\n

But I do think a few rules still make sense:

\n\n

Also, physical media is not just better for preservation. It is often better for your wallet. You can buy a used game, finish it, and sell it to the next guy. Try doing that with a PSN license.

\n

Part 2 is going deeper into software subscriptions specifically, because that rabbit hole gets even dumber. Fitness apps, Adobe, “AI features”, apps charging monthly fees to write text into SQLite, all the usual modern nonsense. I already touched that nerve a bit in my Musclog post about nutrition tracking and scammy fitness app subscriptions, but I have more to say.

\n

For now, that’s the whole rant: cloud stuff is convenient, subscriptions are sometimes unavoidable, and ownership still matters. Especially when the companies telling you it doesn’t are the exact ones with a financial incentive to make sure you never really keep anything again.

\n

See you in part 2.

","otherLanguagesUrl":["/pt-br/blog/general/voce-nao-tera-nada-e-sera-feliz-parte-1/"],"rss":{"title":"You'll own nothing and you'll be happy - Part 1","description":"You'll own nothing and you'll be happy - Part 1","date":"2026-05-16T00:00:00.000Z"},"images":[null],"videos":[],"comments":[],"googleFormData":{"fvv":1,"pageHistory":0,"fbzx":"8917922801126273338","action":"e/1FAIpQLSeST_0jBnLKkEkXGpyx9LWrvV2a1-1F5dr-AcA4wn0BSRyPCw","title":"blog_comment","description":null,"fields":[{"label":"name","description":null,"type":"SHORT_ANSWER","id":"1953327618","required":true},{"label":"email","description":null,"type":"SHORT_ANSWER","id":"1309141965","required":false},{"label":"twitter","description":null,"type":"SHORT_ANSWER","id":"740186305","required":false},{"label":"comment","description":null,"type":"LONG_ANSWER","id":"1663940054","required":true},{"label":"post_path","description":null,"type":"SHORT_ANSWER","id":"1852628638","required":true}],"fieldsOrder":{"740186305":2,"1309141965":1,"1663940054":3,"1852628638":4,"1953327618":0}},"pageType":"blogPost","categoryImage":"/categories_general.jpg/","slug":"/2026/2026-05-16-youll-own-nothing-and-youll-be-happy-part-1.en/","locale":"en","title":"You'll own nothing and you'll be happy - Part 1","previous":{"excerpt":"Uma das side quests mais chatas de publicar um app na Google…","html":"

Uma das side quests mais chatas de publicar um app na Google Play Store é que o Google às vezes decide que você também virou dono de site.

\n

Lá em 2024, quando eu escrevi pela primeira vez sobre o Musclog, o site existia basicamente porque o Google queria que ele existisse. Eu precisava da política de privacidade, das páginas públicas, do kit básico do desenvolvedor de app respeitável, então subi rapidinho um site pequeno em Next.js, consegui aprovar o app, e segui com a minha vida.

\n

Funcionou perfeitamente até o dia em que eu precisei mexer no site de novo.

\n

Screenshots novas? Outro repo. Texto de feature nova? Outro repo. Atualização de página legal? Outro repo. Ajustezinho de tradução? Outro repo.

\n

Nada disso era difícil de verdade, o que quase piorava a situação. Era só chato o suficiente pra continuar me lembrando que eu tinha dividido um produto em duas bases de código por um motivo que eu já não respeitava mais.

\n

Então agora eu não faço mais isso.

\n

O site do Musclog mora dentro do repo do app. Mesmo projeto com Expo Router. Mesmo deploy. musclog.app é o site público, musclog.app/app é o app web de verdade, e Android + iOS continuam saindo da mesma base de código.

\n

Boa prática? Questionável. Conveniente? Demais.

\n

\n \n ;\r\nimport { useEffect } from 'react';\r\nimport { Platform } from 'react-native';\r\n\r\nexport default function Index() {\r\n const router = useRouter();\r\n\r\n useEffect(() => {\r\n if (Platform.OS === 'web') {\r\n router.replace('/home');\r\n }\r\n }, [router]);\r\n\r\n if (Platform.OS === 'web') {\r\n return null;\r\n }\r\n\r\n return <Redirect href=\"/app\" />;\r\n}\n

É isso. No web, / vira o site. No native, / vira o app.

\n

O site também ganha o próprio layout mais leve:

\n
// app/(website)/_layout.web.tsx\r\nimport { Slot } from 'expo-router';\r\n\r\nimport { WebsiteChrome } from '@/components/website/WebsiteChrome';\r\nimport { WebsiteProviders } from '@/components/website/WebsiteProviders';\r\n\r\nexport default function WebsiteLayout() {\r\n  return (\r\n    <WebsiteProviders>\r\n      <WebsiteChrome>\r\n        <Slot />\r\n      </WebsiteChrome>\r\n    </WebsiteProviders>\r\n  );\r\n}
\n

Isso importa porque o layout do app de verdade já tinha crescido e virado um layout de app mesmo. Boot de banco, migrações, React Query, safe area, gestos, modais, snackbars, câmera, contexto do coach, todas aquelas mentiras clássicas de “é só um projetinho paralelo”. O site não precisa de nada disso só pra explicar o que o app faz.

\n

E se algum usuário nativo cair numa rota exclusiva do site, a correção é maravilhosamente direta:

\n
// app/(website)/home.tsx\r\nimport { Redirect } from 'expo-router';\r\n\r\nexport default function Home() {\r\n  return <Redirect href=\"/app\" />;\r\n}
\n

Pronto. Volta pros macros.

\n

\n \n \n \n \n
Site público do Musclog renderizado pela mesma base de código com Expo Router do app
\n

\n

A palhaçada do celular no desktop

\n

Foi aqui que eu parei de me comportar como uma pessoa normal.

\n

Eu queria que musclog.app/app, no desktop, mostrasse o app funcionando dentro de uma moldura de celular. Um pouco porque fica bonito. Um pouco porque funciona como preview do produto. Mas principalmente porque, depois que a ideia entrou na minha cabeça, qualquer opção menos ridícula começou a parecer errada.

\n

O detalhe chato era que /app precisava parecer um celular, mas /home definitivamente não. Então só o roteamento não bastava. Eu precisava que o shell bruto de HTML soubesse, antes da hidratação, se o navegador deveria renderizar a gambiarra do celular fake ou não.

\n

Essa lógica mora em app/+html.tsx, dentro de um script no <head> do documento:

\n
function landingPanelGate(base: string) {\r\n  try {\r\n    function update() {\r\n      const raw = window.location.pathname;\r\n      const path = (base && raw.startsWith(base) ? raw.slice(base.length) : raw) || '/';\r\n\r\n      if (!path.startsWith('/app')) {\r\n        document.documentElement.classList.add('hide-desktop-wrapper');\r\n      } else {\r\n        document.documentElement.classList.remove('hide-desktop-wrapper');\r\n      }\r\n    }\r\n\r\n    update();\r\n    window.addEventListener('popstate', update);\r\n\r\n    const origPush = history.pushState.bind(history);\r\n    history.pushState = function (...args) {\r\n      origPush(...args);\r\n      update();\r\n    };\r\n\r\n    const origReplace = history.replaceState.bind(history);\r\n    history.replaceState = function (...args) {\r\n      origReplace(...args);\r\n      update();\r\n    };\r\n  } catch (_) {}\r\n}
\n

A parte importante é a gambiarra em pushState / replaceState. A primeira renderização foi tranquila. A navegação client-side é que era a parte chata. Sem isso, dava pra sair de /app e continuar com o shell errado ali, como se o Chrome tivesse esquecido em que página estava.

\n

O shell HTML em si é basicamente o painel da landing, o app roteado, e a moldura do celular:

\n
<body className=\"expo-web-body\">\r\n  <div className=\"expo-web-landing\">...</div>\r\n  <script dangerouslySetInnerHTML={{ __html: LANDING_I18N_SCRIPT }} />\r\n  <div className=\"expo-web-root\">\r\n    <div className=\"expo-web-app-shell\">{children}</div>\r\n    <img\r\n      className=\"expo-web-phone-frame\"\r\n      src={withExpoBaseUrl(PHONE_FRAME_SRC)}\r\n      alt=\"\"\r\n      aria-hidden\r\n    />\r\n  </div>\r\n</body>
\n

Aí o CSS comete o crime:

\n
@media (min-width: 1024px) {\r\n  .expo-web-root {\r\n    --frame-h: min(100dvh, max(min(360px, 100dvh), 85dvh));\r\n    aspect-ratio: 1438 / 2976;\r\n    width: min(100vw, calc(var(--frame-h) * 1438 / 2976));\r\n    max-height: var(--frame-h);\r\n    overflow: hidden;\r\n  }\r\n\r\n  .expo-web-app-shell {\r\n    position: absolute;\r\n    left: 7.4409%;\r\n    top: 2.9906%;\r\n    right: 6.3282%;\r\n    bottom: 3.125%;\r\n    zoom: 0.85;\r\n  }\r\n}
\n

Sim, eu medi a área transparente da tela dentro do PNG até o app encaixar direito. Não tem geometria elegante nenhuma por trás dessas porcentagens. Era só eu, uma moldura de celular, e muito empurra-pra-lá até o app parar de parecer torto.

\n

Nas rotas que não são /app, o espetáculo inteiro é desligado:

\n
.hide-desktop-wrapper .expo-web-landing,\r\n.hide-desktop-wrapper .expo-web-phone-frame {\r\n  display: none !important;\r\n}\r\n\r\n.hide-desktop-wrapper .expo-web-app-shell {\r\n  position: static !important;\r\n  width: 100% !important;\r\n  height: auto !important;\r\n  zoom: 1 !important;\r\n  overflow: visible !important;\r\n}
\n

Sim, usa !important. Isso é CSS de pré-hidratação cujo trabalho inteiro é manter a mentira de pé. Não estamos fazendo arquitetura refinada nessa camada.

\n

\n \n \n \n \n
App web do Musclog no desktop renderizado dentro de uma moldura de celular ao lado da landing
\n

\n

Três pequenos crimes

\n

\"Direto

\n

Depois que o shell funcionou, os incômodos menores começaram a aparecer como se tivessem marcado horário.

\n

1. O painel da landing precisava de i18n antes da hidratação

\n

O texto do lado do celular mora em HTML cru, o que significa que React e i18n ainda estão dormindo quando a página renderiza pela primeira vez. Então, se eu quisesse que a página em português não desse um flash de inglês antes, eu tinha que remendar isso manualmente a partir do localStorage:

\n
function landingI18nPatcher(translations, storageKey) {\r\n  try {\r\n    let lang = localStorage.getItem(storageKey);\r\n    let s = (lang && translations[lang]) || translations['en-US'];\r\n\r\n    document.querySelectorAll('[data-landing-i18n]').forEach(function (el) {\r\n      let k = el.getAttribute('data-landing-i18n');\r\n      if (k && s[k]) {\r\n        el.textContent = s[k];\r\n      }\r\n    });\r\n  } catch (_) {}\r\n}
\n

Não é elegante. Também evita jogar um flashbang de texto em inglês numa página em português, então eu vou manter.

\n

2. Assets em HTML cru não recebem a ajuda de sempre do Expo

\n

O PNG da moldura do celular e o QR code também são referenciados a partir de HTML cru, então aquela conveniência normal de assets do Expo não me ajudava muito ali. Resultado: eu precisei de um helper pra base URL:

\n
function withExpoBaseUrl(path: string): string {\r\n  const base = process.env.EXPO_BASE_URL;\r\n  if (base == null || base === '') {\r\n    return path;\r\n  }\r\n\r\n  const basePath = String(base).replace(/^\\/+|\\/+$/g, '');\r\n  const normalized = path.startsWith('/') ? path : `/${path}`;\r\n  return `/${basePath}${normalized}`;\r\n}
\n

E como esses arquivos vivem fora do pipeline normal de assets do React Native, eu copio tudo pra public/ antes de rodar dev e export. Esquece esse passo uma vez e o site imediatamente te lembra quem manda.

\n

3. O expo export às vezes precisava de apoio emocional

\n

Em alguns ambientes, expo export --platform web terminava com sucesso e depois só… continuava vivo sem motivo nenhum. Pasta dist lá. Arquivos gerados. Processo espiritualmente concluído, tecnicamente ainda pendurado.

\n

Então agora existe um script wrapper:

\n
// scripts/export-web-wrapper.js\r\nif (output.includes('Exported: dist')) {\r\n  console.log('[export-web-wrapper] Detected successful export. Forcing exit in 5s...');\r\n  setTimeout(() => process.exit(0), 5000);\r\n}
\n

Não é glamouroso… mas resolve.

\n

Aí os modais ficaram esquisitos

\n

\n \n \n \n \n
Modal do Musclog no desktop renderizado corretamente dentro da moldura do celular
\n

\n

Claro que ficaram.

\n

Quando o app web mora dentro de um celular fake no desktop, o comportamento padrão de Modal no React Native Web começa a ficar ridículo bem rápido. Fazer portal direto pro document.body funciona bem quando o app é dono da página inteira. Funciona bem menos quando o app está visualmente recortado dentro de uma carcaça de celular e o modal decide do nada que agora pertence à janela inteira do navegador.

\n

Então eu adicionei um WebModalShellProvider lá em cima em app/app/_layout.tsx, com um host de overlay dentro do shell do celular:

\n
// context/WebModalShellContext.web.tsx\r\nreturn (\r\n  <WebModalShellContext.Provider value={{ hostElement }}>\r\n    <View style={{ flex: 1, minHeight: 0, height: '100%', width: '100%', position: 'relative' }}>\r\n      <View style={{ flex: 1, minHeight: 0, height: '100%', width: '100%' }} collapsable={false}>\r\n        {children}\r\n      </View>\r\n      <View\r\n        id=\"expo-web-modal-shell-host\"\r\n        ref={setHostRef}\r\n        collapsable={false}\r\n        style={{\r\n          position: 'absolute',\r\n          left: 0,\r\n          right: 0,\r\n          top: 0,\r\n          bottom: 0,\r\n          zIndex: 1_000_000,\r\n        }}\r\n      />\r\n    </View>\r\n  </WebModalShellContext.Provider>\r\n);
\n

Aí o Modal.web.tsx faz aquela coisa óbvia que só parece óbvia depois, alternando entre um portal dentro desse host e o modal normal do RN:

\n
const useShellPortal = Platform.OS === 'web' && isDesktopFrame && hostElement != null;\r\n\r\nif (useShellPortal) {\r\n  return createPortal(\r\n    <View\r\n      style={{\r\n        position: 'absolute',\r\n        left: 0,\r\n        right: 0,\r\n        top: 0,\r\n        bottom: 0,\r\n        width: '100%',\r\n        height: '100%',\r\n      }}\r\n    >\r\n      {children}\r\n    </View>,\r\n    hostElement\r\n  );\r\n}\r\n\r\nreturn <RNModal visible={visible}>{children}</RNModal>;
\n

E o hook compartilhado do overlay sabe se ele deve se comportar como um modal de viewport inteira ou como um modal de viewport de celular fake:

\n
export function useWebModalLayerStyle(options = {}) {\r\n  const isDesktopFrame = useWebDesktopPhoneFrame();\r\n\r\n  if (isDesktopFrame) {\r\n    return {\r\n      position: 'absolute',\r\n      top: 0,\r\n      left: 0,\r\n      right: 0,\r\n      bottom: 0,\r\n      width: '100%',\r\n      height: '100%',\r\n    };\r\n  }\r\n\r\n  return {\r\n    position: 'fixed',\r\n    top: 0,\r\n    left: 0,\r\n    right: 0,\r\n    bottom: 0,\r\n    width: '100vw',\r\n    height: '100dvh',\r\n  };\r\n}
\n

Isso resolveu a parte visual. Aí os pointer events decidiram que também queriam atenção.

\n

No native, pointerEvents=\"box-none\" é normal. No HTML, isso vira pointer-events: box-none, que não é CSS de verdade, então o navegador fica livre pra improvisar. Então agora existe uma regra defensiva pro host do overlay:

\n
.expo-web-app-shell #expo-web-modal-shell-host {\r\n  pointer-events: none !important;\r\n}\r\n\r\n.expo-web-app-shell #expo-web-modal-shell-host * {\r\n  pointer-events: auto !important;\r\n}
\n

Também existe um useLayoutEffect forçando a mesma coisa inline, porque a essa altura eu já não tinha mais interesse em descobrir quão elegante seria a solução correta.

\n

Eu claramente tenho um padrão

\n

Se você já lê este blog há algum tempo, nada disso deveria te surpreender.

\n

Eu já transformei este blog num RPG top-down porque o Konami Code merecia uma recompensa maior do que só um efeitinho de fundo.

\n

Depois eu coloquei Stories no blog porque o app da minha operadora tinha Stories, e isso me irritou profundamente num nível suficiente pra eu transformar o problema em problema de todo mundo.

\n

Então sim, é claro que agora o meu app de fitness serve o site a partir do mesmo repo com Expo Router, e é claro que o app web no desktop roda dentro de uma foto gigante de um celular. Isso não é uma escalada surpreendente. É só continuidade.

\n

Por que eu vou manter isso

\n

Passando da piada, isso resolveu um imposto de manutenção bem real.

\n

As traduções são compartilhadas. Os design tokens são compartilhados. Os componentes são compartilhados. As páginas legais ficam do lado do produto que elas estão legalmente defendendo. Quando eu mudo texto, screenshot ou posicionamento de feature, eu não preciso lembrar qual repo guarda a versão pública e comportada do mesmo app.

\n

O deploy também ficou melhor. Um build. Um export. Um artefato. O site e o app web não têm como sair de sincronia a não ser que eu faça questão de separá-los, o que fica bem mais difícil quando eles literalmente são publicados juntos.

\n

E quando eu lancei o redesign do Musclog, o site basicamente veio junto no embalo. Mesmo spacing, mesmas cores, screenshots atualizadas, mesmo PR. Exatamente o tipo de vitória chata de manutenção que faz uma decisão de arquitetura meio estranha parecer inteligente seis meses depois.

\n

Conclusão

\n

Eu poderia ter mantido o site num repo separado em Next.js como uma pessoa emocionalmente mais estável? Poderia.

\n

Eu ia continuar fazendo isso depois que o Expo Router deixou esse caminho aberto? Não.

\n

A parte engraçada é que o roteamento foi a parte sem graça. O Expo Router resolveu isso muito bem. O trabalho irritante estava em todas as mentirinhas de navegador ao redor: bloqueio de rota antes da hidratação, i18n em HTML cru, sincronização de assets, modais que sabiam em que shell estavam, esquisitice com pointer-events, e convencer o Chrome do desktop a acreditar temporariamente que era um celular.

\n

Ainda assim, valeu muito a pena.

\n

Se quiser fuçar o código, o Musclog é open-source no GitHub. Se quiser usar, tá em musclog.app. O site mora dentro do repo do app agora.

\n

Agora mora todo mundo aqui.

","fields":{"postHashId":"Y29kaW5ndHJ1ZW51bGwyMDI2LTA0LTI1VDAwOjAwOjAwLjAwMFo=","slug":"/2026/2026-04-25-site-ou-app-sim-o-guia-gambiarrístico-para-sites-feitos-em-expo-router.pt-br/","path":"/blog/coding/site-ou-app-sim-o-guia-gambiarrístico-para-sites-feitos-em-expo-router/","locale":"pt-br"},"rawMarkdownBody":"\r\nUma das side quests mais chatas de publicar um app na Google Play Store é que o Google às vezes decide que você também virou dono de site.\r\n\r\nLá em 2024, quando eu [escrevi pela primeira vez sobre o Musclog](/pt-br/blog/coding/musclog-aproveitando-minha-experiencia-com-reactjs-para-criar-um-app-em-react-native/), o site existia basicamente porque o Google queria que ele existisse. Eu precisava da política de privacidade, das páginas públicas, do kit básico do desenvolvedor de app respeitável, então subi rapidinho um site pequeno em Next.js, consegui aprovar o app, e segui com a minha vida.\r\n\r\nFuncionou perfeitamente até o dia em que eu precisei mexer no site de novo.\r\n\r\nScreenshots novas? Outro repo. Texto de feature nova? Outro repo. Atualização de página legal? Outro repo. Ajustezinho de tradução? Outro repo.\r\n\r\nNada disso era difícil de verdade, o que quase piorava a situação. Era só chato o suficiente pra continuar me lembrando que eu tinha dividido um produto em duas bases de código por um motivo que eu já não respeitava mais.\r\n\r\nEntão agora eu não faço mais isso.\r\n\r\nO site do Musclog mora dentro do repo do app. Mesmo projeto com Expo Router. Mesmo deploy. [musclog.app](https://musclog.app/) é o site público, [musclog.app/app](https://musclog.app/app) é o app web de verdade, e Android + iOS continuam saindo da mesma base de código.\r\n\r\nBoa prática? Questionável. Conveniente? Demais.\r\n\r\n![Site público do Musclog renderizado pela mesma base de código com Expo Router do app](../../uploads/blog/2026/04/musclog-shared-router-website-home-1.png)\r\n\r\n## Ok mas por quê?\r\n\r\nBasicamente porque o Expo Router arrancou a minha última desculpa.\r\n\r\nO Musclog já usava [Expo Router](https://expo.github.io/router/docs), então o app já estava organizado em rotas baseadas em arquivos dentro da pasta `app/`. Aí eu percebi que podia simplesmente criar um grupo de rotas chamado `(website)`, mover as páginas públicas pra lá, deixar o app de verdade em `app/app/*`, e parar de fingir que eu estava lidando com dois produtos diferentes.\r\n\r\nO repo separado tinha virado uma daquelas decisões de arquitetura que parecem limpas na teoria e depois ficam te cobrando imposto de manutenção pra sempre:\r\n\r\n- dois PRs pra uma mudança num produto\r\n- dois deploys\r\n- dois lugares pras traduções\r\n- dois lugares pras páginas legais\r\n- dois lugares pra esquecer alguma coisa\r\n\r\nSem condições.\r\n\r\n## A divisão\r\n\r\nA parte engraçada é que o roteamento em si acabou sendo a parte menos amaldiçoada dessa história:\r\n\r\n- `app/app/*` é o Musclog de verdade. Treinos, nutrição, coach com IA, tudo.\r\n- `app/(website)/*` é o site público. Landing page, política de privacidade, termos, contato, calculadora.\r\n\r\nA rota raiz só checa a plataforma e te joga pra onde você pertence:\r\n\r\n```tsx\r\n// app/index.tsx\r\nimport { Redirect, useRouter } from 'expo-router';\r\nimport { useEffect } from 'react';\r\nimport { Platform } from 'react-native';\r\n\r\nexport default function Index() {\r\n const router = useRouter();\r\n\r\n useEffect(() => {\r\n if (Platform.OS === 'web') {\r\n router.replace('/home');\r\n }\r\n }, [router]);\r\n\r\n if (Platform.OS === 'web') {\r\n return null;\r\n }\r\n\r\n return ;\r\n}\r\n```\r\n\r\nÉ isso. No web, `/` vira o site. No native, `/` vira o app.\r\n\r\nO site também ganha o próprio layout mais leve:\r\n\r\n```tsx\r\n// app/(website)/_layout.web.tsx\r\nimport { Slot } from 'expo-router';\r\n\r\nimport { WebsiteChrome } from '@/components/website/WebsiteChrome';\r\nimport { WebsiteProviders } from '@/components/website/WebsiteProviders';\r\n\r\nexport default function WebsiteLayout() {\r\n return (\r\n \r\n \r\n \r\n \r\n \r\n );\r\n}\r\n```\r\n\r\nIsso importa porque o layout do app de verdade já tinha crescido e virado um layout de app mesmo. Boot de banco, migrações, React Query, safe area, gestos, modais, snackbars, câmera, contexto do coach, todas aquelas mentiras clássicas de \"é só um projetinho paralelo\". O site não precisa de nada disso só pra explicar o que o app faz.\r\n\r\nE se algum usuário nativo cair numa rota exclusiva do site, a correção é maravilhosamente direta:\r\n\r\n```tsx\r\n// app/(website)/home.tsx\r\nimport { Redirect } from 'expo-router';\r\n\r\nexport default function Home() {\r\n return ;\r\n}\r\n```\r\n\r\nPronto. Volta pros macros.\r\n\r\n![Site público do Musclog renderizado pela mesma base de código com Expo Router do app](../../uploads/blog/2026/04/musclog-shared-router-website-home-2.png)\r\n\r\n## A palhaçada do celular no desktop\r\n\r\nFoi aqui que eu parei de me comportar como uma pessoa normal.\r\n\r\nEu queria que [musclog.app/app](https://musclog.app/app/), no desktop, mostrasse o app funcionando dentro de uma moldura de celular. Um pouco porque fica bonito. Um pouco porque funciona como preview do produto. Mas principalmente porque, depois que a ideia entrou na minha cabeça, qualquer opção menos ridícula começou a parecer errada.\r\n\r\nO detalhe chato era que `/app` precisava parecer um celular, mas `/home` definitivamente não. Então só o roteamento não bastava. Eu precisava que o shell bruto de HTML soubesse, antes da hidratação, se o navegador deveria renderizar a gambiarra do celular fake ou não.\r\n\r\nEssa lógica mora em `app/+html.tsx`, dentro de um script no `` do documento:\r\n\r\n```tsx\r\nfunction landingPanelGate(base: string) {\r\n try {\r\n function update() {\r\n const raw = window.location.pathname;\r\n const path = (base && raw.startsWith(base) ? raw.slice(base.length) : raw) || '/';\r\n\r\n if (!path.startsWith('/app')) {\r\n document.documentElement.classList.add('hide-desktop-wrapper');\r\n } else {\r\n document.documentElement.classList.remove('hide-desktop-wrapper');\r\n }\r\n }\r\n\r\n update();\r\n window.addEventListener('popstate', update);\r\n\r\n const origPush = history.pushState.bind(history);\r\n history.pushState = function (...args) {\r\n origPush(...args);\r\n update();\r\n };\r\n\r\n const origReplace = history.replaceState.bind(history);\r\n history.replaceState = function (...args) {\r\n origReplace(...args);\r\n update();\r\n };\r\n } catch (_) {}\r\n}\r\n```\r\n\r\nA parte importante é a gambiarra em `pushState` / `replaceState`. A primeira renderização foi tranquila. A navegação client-side é que era a parte chata. Sem isso, dava pra sair de `/app` e continuar com o shell errado ali, como se o Chrome tivesse esquecido em que página estava.\r\n\r\nO shell HTML em si é basicamente o painel da landing, o app roteado, e a moldura do celular:\r\n\r\n```tsx\r\n\r\n
...
\r\n