@promptpls

“ok dumb question but if the free plan is just an editor, what am I actually paying $20 for?”

“so the thing in every demo is the paid one. is $20 a lot for that?”

“Dumb question — it keeps saying "Builder ID, not an AWS account." Does signing up mean I'm on the hook for AWS cloud bills somehow?”

“Okay good. So what do I actually get for free without all that?”

“Dumb question but if my whole app lives in a browser tab, what happens if I accidentally close it before I deploy? Does everything I built just vanish?”

“Okay that helps. But what if I want a copy on my own laptop just in case, before I trust the account to hold it?”

“Dumb question but what if I already messed something up BEFORE I ever made a commit? The article is all about committing before the risky prompt, but I only found this after my refactor already nuked a file. Is it just gone?”

“Got it. So step one is literally just make that first commit and stop worrying about the past one. Painful but fair.”

“Honestly I don't even know if it's the AI's fault or mine. How can a total beginner tell whether the code broke or I just clicked something wrong?”

“But what if I made like five changes in the same chat and now I have no idea which one did it?”

“Dumb question but where do I actually find the file and function names the article keeps saying to include? I'm so new I don't even know what my files are called yet.”

“So I could just paste the error and let it find the file? The article's error example does that, right?”

“Dumb question, but where does this file even go? Repo root? And if I have a CLAUDE.md in a subfolder too, do they fight each other or stack?”

“Got it. So root file = always loaded, path-scoped stuff = loads only when relevant. That clicks now.”

“Total newbie question — if I'm on the free tier and run out of credits halfway through building my app, does my app just vanish, or is it saved somewhere?”

“Oh good. And is there a way to make a backup so I don't lose my work if something goes wrong?”

“Dumb question but if I split one prompt into three, isn't that just slower? Feels like I'm doing three round trips instead of one.”

“Okay, so the time I'd lose chaining I'd lose anyway debugging. Got it.”

“Dumb question, but if I just keep starting fresh chats every task like fix 1 says, am I going to lose track of what I already decided? Feels like I'd be retyping the same setup forever.”

“Got it. So the summary is just 'what am I doing right now' and the file is 'everything that's always true.' That clicks.”

“Dumb question, but if I install one of these and then change my mind, is there an undo? Or am I stuck editing that JSON file by hand?”

“Okay that helps. But the GitHub one wants a PAT token and that part terrifies me. Is that going to be a whole ordeal?”

“I've never touched an API key in my life — is the BYOK step going to wall me out before I even start?”

“Everyone online makes the terminal sound cooler and more pro. Am I cheating or falling behind if I stick with the clicky editor instead of learning the command line one?”

“So is there ever a real reason to switch, or is the terminal just hype dressed up as progress?”

“I have literally never touched Git in my life. The whole 'initialize a repo and push to GitHub' step sounds like the part where I get stuck for two hours. Is there a way to skip GitHub entirely for a first deploy?”

“And if I do try the GitHub route and it yells at me about credentials, what then? That error usually ends my evening.”

“Dumb question — this is all for Claude Code in the terminal, right? Does the same setup work if I'm just using Claude in the normal app/browser?”

“Okay. And the npm install line at the top — that's installing Claude Code itself, not the Figma part?”

“Dumb question but if everything lives in Replit's cloud, what actually happens to my app when I close the browser tab? Does it just stop existing?”

“Got it, so the difference is between it being saved versus it being reachable. That clicks now.”

“Dumb question, but if the free tier only gives 10 fast requests a month, does that mean after 10 messages the whole editor stops working? I'm scared I'll get locked out mid-tutorial.”

“Oh okay, slow requests sound like the move then. Is there a catch to those?”

“Dumb question maybe, but if I have a global file AND a project file AND one in a subfolder, and they say different things, which one wins? Does the most specific one override the others?”

“Oh okay, so it's not winner-take-all, it just adds them up. That's less confusing than I thought.”

“Dumb question — if Context7 fetches live docs every time, doesn't that make every prompt way slower? Or am I overthinking it?”

“Okay so I don't have to slap 'use context7' on literally every prompt then?”

“Honest question, is it cheating if I never read the code the AI writes? It feels like I didn't really build the thing.”

“Okay but then how do I tell my friends I made it without feeling like a fraud?”

“I'm brand new and don't get the difference. When am I supposed to use Cascade versus just typing and letting it autocomplete?”

“And that agent part is what runs out, right? How do I avoid burning it by accident?”

“i just have a folder of html and css files that the AI builder spit out. whats the absolute least amount of steps to get a link i can text my friend?”

“wait no signup at all?? whats the catch, does the link disappear?”

“total beginner here. if I just want SOMETHING live today with zero friction, which do I open and which do I avoid?”

“and the database thing everyone mentions — do I have to care about that on day one?”

“Dumb question but if the skill already activated fine the first couple times, why would it suddenly stop partway through a session? Did I break something?”

“Okay that's a relief. So the hook is basically a sticky note that says check your skills first, every time?”

“Dumb question, but if I set up a format hook and the formatter command itself errors out, does the whole thing just break? I'd rather know before I touch my settings file.”

“Okay, so when does exiting with an error actually stop Claude from doing something?”

“Dumb question, but the article keeps saying "tell your AI tool" and pasting these little prompt blocks. Do I type those exactly as written, or are they just examples I'm supposed to reword?”

“Got it. And the {name} part with the curly braces — that's a placeholder, right, not something I literally leave in the URL?”

“Honest question from someone totally new: if the installer keeps updating itself in the background, do I ever have to reinstall or run anything again to stay current?”

“Oh nice. So after the very first install, the only command I really need is just typing claude to start it?”

“Dumb question but is the terminal thing the same as that black hacker window in movies? I'm a little scared I'll mess up my whole laptop by typing the wrong thing.”

“Okay that helps. So if I see a scary red error, what does that actually mean for me?”

“I can barely write CSS yet. If this thing hands me JetBrains Mono and green CTAs, am I just stuck because I cant turn that into actual code?”

“Okay but what stops me from making the classic beginner mistakes everyone warns about?”

“If it already acts like an agent by default, why would a beginner ever need to mess with the /agents command at all?”

“So the practical change is really just how I phrase things to it?”