Filter
Exclude
Time range
-
Near
Replying to @toddgeist @swyx
True. In @QwikBuild, we made a lot of opinionated choices for all of these, ensuring that users get a fully working app, after multiple hours of agent execution.
4
Replying to @swyx
The "webmaster infra setup" is pretty much solved for apps made on @QwikBuild because we took an opinionated approach. "Once and for all" would be nice though.
363
@QwikBuild Look at that Cache Hit Rate. I'm impressed.
1
15
I almost paid ₹3,799/year for unlimited chess puzzles on @chesscom Instead, I built my own chess puzzle app with @QwikBuild I’m on chess .com almost every day. It’s my default time killer, and chess puzzles are the perfect tiny dopamine hit. But the daily limits were annoying, and I also wanted a practice flow that matched how I actually improve. Time to summon my other addiction: Building personal software. So I opened Qwikbuild’s console and typed: “I want an app to practice chess. given a starting position the puzzle should be about finding the best move or find mate in 1 or 2.” AgentQ then: 1. Asked me a few clarifying questions 2. Generated a PRD that I could review 3. Asked me to recheck the implementation plan 4. Boom...built the app and deployed it No headache of how to run it, how to version it, where to host it, or how to deploy it, etc. Just an idea turned into a working web app. I made a few small tweaks after that, but the core flow was done end to end. Numbers: Tokens: 17M AI cost: Rs. 2,547 / $26.72 Time: 1h 32m Around 1 hour was just LLM calls. The rest was command execution and deployment pipelines. This is exactly the kind of personal software I think more people will build. Small, personal, useful apps that solve one annoying problem. Whether you’re an individual or an organization, stop being limited by cookie-cutter software in big 2026. Make your own cookies with @QwikBuild
1
1
6
168
Agents on their own are not that useful. A standalone ticket-booking agent, for instance. You'd use it once in a while when you remember it exists. No pull to come back because it knows nothing about your work. Now take an agent that sits on top of an app you've already built. That app has your business logic, your workflows, your use case. Everything you spent time describing and configuring. That's where your context lives. The problem is that context stays locked inside the deployed web app. When you're in email, Slack, or WhatsApp, it doesn't follow you there. Say you're an HR and you've built an applicant tracking app. Filters, reminders, candidate pipelines, all configured. But candidates reach you through cold emails, WhatsApp messages with attached resumes. Every time, you switch to the app, add details manually, switch back. Works, but the friction adds up fast. An agent layer that knows your app's context changes this. You stay where you are, say "got a new candidate, here's the resume," and it handles the rest. The context comes to you instead of you going to it. @QwikBuild will solve this soon, stay tuned...
1
4
59
The current state of LLMs and AI reminds me of lines spoken in Person of Interest series: “Don’t tie your life to its whims. We cannot understand these intelligences. The best we can hope for is to survive them.” That line was written over a decade ago about “The Machine” — a fictional superintelligence that watched everything, predicted everything, and eventually slipped beyond human control. Today? It doesn’t feel fictional anymore. We’re not just using LLMs. We’re building entire careers, companies, creative processes, and even daily decision-making around them. Prompt chains. Agent swarms. Autonomous workflows. Memory layers. Fine-tuned models that know us better than our closest colleagues. Dependency has grown exponentially. What started as “let me just get a second opinion” has quietly become “I can’t ship anything without running it through the model first.” And here’s the part that should actually concern you: This is exactly what the model companies want us to become. They don’t need us to love the technology. They need us to be unable to live without it. Every new feature, every context window increase, every “agent” upgrade is engineered to make the tool feel indispensable. The more we outsource our thinking, our memory, our judgment — the deeper the lock-in becomes. We feed them our data, our workflows, our edge cases… and in return we get faster, smarter, more addictive assistance. Until one day we wake up and realize we no longer know how to operate without it. We cannot understand these intelligences. Not really. They are black boxes trained on incomprehensible scales of data, evolving at a pace no human team can match. Their “whims” — the random hallucinations, the subtle biases, the sudden capability jumps, the silent policy changes — are already shaping how millions of people think and work. So maybe fictional character Finch was right. Don’t tie your life to its whims. Use the tools. Master them. But keep your own compass. Build systems that augment you — not ones that quietly replace your judgment. Keep skills the models can’t touch. Protect the parts of your mind that still belong only to you. Because survival isn’t about rejecting the technology. It’s about refusing to become dependent on something we were never meant to fully understand in the first place. The Machine was fiction. What we’re building now isn’t. Stay vigilant. #PersonOfInterest #AI #LLM #Superintelligence #TheMachine #Claudecode #anthropic #openai #gemini #grok #qwikbuild #agentq #ai
2
1
5
115
In the current times of vulnerabilities popping up everywhere, this is much needed. Brought to you by @QwikBuild
You can now scan your @qwikbuild projects for any malware, vulnerabilities or license violations in your package dependencies.
3
73
In the age of AI, every investment in security will compound. We take security seriously at @QwikBuild
You can now scan your @qwikbuild projects for any malware, vulnerabilities or license violations in your package dependencies.
1
3
240
You can now scan your @qwikbuild projects for any malware, vulnerabilities or license violations in your package dependencies.
3
5
40
950
We solved the hard problems of making sure the our agents could handle changes while building the Qwikbuild platform, but how users made those changes is what visual feedback is for. Love how easy it is to make changes to my apps as and when I discover new things.
Introducing Visual Feedback on @QwikBuild The number one PAIN POINT across customers: giving change requests to AgentQ is harder than it should be. Most users - developers, designers, and non-technical people alike - struggle to explain WHERE something is and WHAT needs to change. Not everyone thinks in terms like sticky headers, modals, hamburger menus, or layout containers. So feedback turns into long descriptions, screenshots, and back-and-forth clarification: “On the page after I click A, where the text B appears under section C, change it to D.” That is too much friction. And it leaves customers exhausted. Visual Feedback makes this natural. Open the live preview next to your chat, click the exact area you want to change, and leave a comment. The agent gets the visual context, the element context, and your instruction. All the complexity of describing the location and element gets consumed by one word: “this” You can still toggle into “Use the app” mode to navigate, sign in, and reach the right state. Once all comments are added, click “Apply Comments,” review them, and send the full context to the agent. Less explaining. Less back-and-forth. Faster changes. Welcome to the bright side.
3
39
Big upgrade!👇 Visual Feedback is here ✨ No more paragraphs on “where” things need to change, Just open the live preview, click the exact element, drop your comment, and let AgentQ see exactly what you mean. Comment 'Feedback' for Qwikbuild access
Introducing Visual Feedback on @QwikBuild The number one PAIN POINT across customers: giving change requests to AgentQ is harder than it should be. Most users - developers, designers, and non-technical people alike - struggle to explain WHERE something is and WHAT needs to change. Not everyone thinks in terms like sticky headers, modals, hamburger menus, or layout containers. So feedback turns into long descriptions, screenshots, and back-and-forth clarification: “On the page after I click A, where the text B appears under section C, change it to D.” That is too much friction. And it leaves customers exhausted. Visual Feedback makes this natural. Open the live preview next to your chat, click the exact area you want to change, and leave a comment. The agent gets the visual context, the element context, and your instruction. All the complexity of describing the location and element gets consumed by one word: “this” You can still toggle into “Use the app” mode to navigate, sign in, and reach the right state. Once all comments are added, click “Apply Comments,” review them, and send the full context to the agent. Less explaining. Less back-and-forth. Faster changes. Welcome to the bright side.
1
5
177
Delivering the no 1 requested feature on @QwikBuild checkout the demo
Introducing Visual Feedback on @QwikBuild The number one PAIN POINT across customers: giving change requests to AgentQ is harder than it should be. Most users - developers, designers, and non-technical people alike - struggle to explain WHERE something is and WHAT needs to change. Not everyone thinks in terms like sticky headers, modals, hamburger menus, or layout containers. So feedback turns into long descriptions, screenshots, and back-and-forth clarification: “On the page after I click A, where the text B appears under section C, change it to D.” That is too much friction. And it leaves customers exhausted. Visual Feedback makes this natural. Open the live preview next to your chat, click the exact area you want to change, and leave a comment. The agent gets the visual context, the element context, and your instruction. All the complexity of describing the location and element gets consumed by one word: “this” You can still toggle into “Use the app” mode to navigate, sign in, and reach the right state. Once all comments are added, click “Apply Comments,” review them, and send the full context to the agent. Less explaining. Less back-and-forth. Faster changes. Welcome to the bright side.
1
6
284
This is huge! Language is not everything. AI agents are much more powerful when they can *see* what you are pointing at. We applied this principle in @QwikBuild for building apps. 😎
Introducing Visual Feedback on @QwikBuild The number one PAIN POINT across customers: giving change requests to AgentQ is harder than it should be. Most users - developers, designers, and non-technical people alike - struggle to explain WHERE something is and WHAT needs to change. Not everyone thinks in terms like sticky headers, modals, hamburger menus, or layout containers. So feedback turns into long descriptions, screenshots, and back-and-forth clarification: “On the page after I click A, where the text B appears under section C, change it to D.” That is too much friction. And it leaves customers exhausted. Visual Feedback makes this natural. Open the live preview next to your chat, click the exact area you want to change, and leave a comment. The agent gets the visual context, the element context, and your instruction. All the complexity of describing the location and element gets consumed by one word: “this” You can still toggle into “Use the app” mode to navigate, sign in, and reach the right state. Once all comments are added, click “Apply Comments,” review them, and send the full context to the agent. Less explaining. Less back-and-forth. Faster changes. Welcome to the bright side.
5
474
Introducing Visual Feedback on @QwikBuild The number one PAIN POINT across customers: giving change requests to AgentQ is harder than it should be. Most users - developers, designers, and non-technical people alike - struggle to explain WHERE something is and WHAT needs to change. Not everyone thinks in terms like sticky headers, modals, hamburger menus, or layout containers. So feedback turns into long descriptions, screenshots, and back-and-forth clarification: “On the page after I click A, where the text B appears under section C, change it to D.” That is too much friction. And it leaves customers exhausted. Visual Feedback makes this natural. Open the live preview next to your chat, click the exact area you want to change, and leave a comment. The agent gets the visual context, the element context, and your instruction. All the complexity of describing the location and element gets consumed by one word: “this” You can still toggle into “Use the app” mode to navigate, sign in, and reach the right state. Once all comments are added, click “Apply Comments,” review them, and send the full context to the agent. Less explaining. Less back-and-forth. Faster changes. Welcome to the bright side.
1
2
6
2,222
Sneak peek into what’s coming next on @QwikBuild
1
3
35
Most vibecoding platforms give you a login and a chatbot. No onboarding. No support. Just a ticket form that goes nowhere. You raise a bug - and wait a week. You ask a question - and get sent to documentation. You get stuck - and the platform just... doesn't care. There's a reason so many ambitious projects die quietly. It's lack of support at the critical moments. We built something different at @QwikBuild : every customer gets a real development team alongside the AI platform. You chat with the AI to build. You escalate to humans when it matters. You ship. That's how it should work. Reach human support here - qwikbuild. com/apply #SaaSDevelopment #DeveloperExperience #WhiteGlove #AIDevOps #ProductBuilding
1
4
8
417
Geofencing daily targets map view of every store visit. Read how Talisva solved field sales pain with @QwikBuild
1
5
241
How @QwikBuild has helped Karnataka’s only commercial fruit wine brand to build custom on-field sales software. 💪🏻 #qwikbuild
2
3
720