We analyzed real conversations across social media to understand how people really feel about Vercel.
Monitoring conversations on
Based on real conversations from social media
How people feel about this brand
See how brands compare based on real sentiment analysis
Real conversations from across the web
AW HELL NAH YO ASS TWEAKIN JIGSAW context: rule34dle is a game where you guess which one has more porn
Does anyone else think the whole "separate database provider" trend is completely backwards? Okay so I'm a developer with 15 years of PHP, NodeJS and am studying for Security+ right now and this is driving me crazy. How did we all just... agree that it's totally fine to host your app on one provider and yeet your database onto a completely different one across the public internet? Examples I have found. * Laravel Cloud connecting to some Postgres instance on Neon (possibly the same one according to other posts) * Vercel apps hitting databases on Neon/PlanetScale/Supabase * Upstash Redis The latency is stupid. Every. Single. Query. has to go across the internet now. Yeah yeah, I know about PoPs and edge locations and all that stuff, but you're still adding a massive amount of latency compared to same-VPC or same-datacenter connections. A query that should take like 1-2ms now takes 20-50ms+ because it's doing a round trip through who knows how many networks. And if you've got an N+1 query problem? Your 100ms page just became 5 seconds. And yes, I KNOW it's TLS encrypted. But you're still exposing your database to the entire internet. Your connection strings all of it is traveling across networks you don't own or control. Like I said, I'm studying Security+ right now and I can't even imagine trying to explain to a compliance/security team why customer data is bouncing through the public internet 50 times per page load. That meeting would be... interesting. Look, I get it - the Developer Experience is stupid easy. Click a button, get a connection string, paste it in your env file, deploy. But we're trading actual performance and security for convenience. We're adding latency, more potential failure points, security holes, and locking ourselves into multiple vendors. All so we can skip learning how to properly set up a database? What happened to keeping your database close to your app? VPC peering? Actually caring about performance? What is everyones thoughts on this?
Vercel CEO meets with Netanyahu to discuss AI education
Infinite World in TypeScript # Try it here [https://generate-infinite.vercel.app/](https://generate-infinite.vercel.app/) Been working on an **immersive, infinite, procedurally generated world** built with **Three.js** and **WebGL**, fully refactored in **TypeScript**. This project is a TypeScript-based evolution of the original [Infinite World](https://github.com/brunosimon/infinite-world) concept by the creative developer Bruno Simon. It serves as: * technical demonstration of procedural generation. * a case study in migrating a JavaScript codebase to TypeScript for enhanced scalability and maintainability. # Key Features * **Infinite Procedural Terrain:** Explores a unique, dynamically generated world every time. * **Built with Three.js:** Leverages the power of **Three.js** for efficient **WebGL** rendering. * **Fully Typed Codebase:** Migrated from JavaScript to **TypeScript** for robust, maintainable, and error-free code. * **Enhanced Rendering:** Features improved lighting, shading, and terrain generation algorithms for a more natural look. * **Collision Detection:** Implemented a camera collision system to prevent clipping through the terrain, improving the user experience. # Notes It was tested on Linux, and it works perfectly well, however some rendering issues have been observed when testing on Windows, it is still a mystery, so your insight and feedback are welcome and appreciated! # More details [ibra-kdbra/Generate\_infinite](https://github.com/ibra-kdbra/Generate_infinite)
Everyone Builds AI Agents. Almost No One Knows How to Deploy Them. I've seen this happen a dozen times with clients. A team spends weeks building a brilliant agent with LangChain or CrewAI. It works flawlessly on their laptop. Then they ask the million-dollar question: "So... how do we get this online so people can actually use it?" The silence is deafening. Most tutorials stop right before the most important part. Your agent is a cool science project until it's live. You can't just keep a terminal window open on your machine forever. So here’s the no nonsense guide to actually getting your agent deployed, based on what works in the real world. The Three Places Your Agent Can Actually Live Forget the complex diagrams. For 99% of projects, you have three real options. * Serverless (The "Start Here" Method): This is the default for most new agents. Platforms like Google Cloud Run, Vercel, or even Genezio let you deploy code directly from GitHub without ever thinking about a server. You just provide your code, and they handle the rest. You pay only when the agent is actively running. This is perfect for simple chatbots, Q&A tools, or basic workflow automations. * Containers (The "It's Getting Serious" Method): This is your next step up. You package your agent and all its dependencies into a Docker container. Think of it as a self-contained box that can run anywhere. You then deploy this container to a service like Cloud Run (which also runs containers), AWS ECS, or Azure Container Apps. You do this when your agent needs more memory, has to run for more than a few minutes (like processing a large document), or has finicky dependencies. * Full Servers (The "Don't Do This Yet" Method): This is managing your own virtual machines or using a complex system like Kubernetes. I'm telling you this so you know to avoid it. Unless you're building a massive, enterprise scale platform with thousands of concurrent users, this is a surefire way to waste months on infrastructure instead of improving your agent. A Dead Simple Path for Your First Deployment Don't overthink it. Here is the fastest way to get your first agent live. 1. Wrap your agent in an API: Your Python script needs a way to receive web requests. Use a simple framework like Flask or FastAPI to create a single API endpoint that triggers your agent. 2. Push your code to GitHub: This is standard practice and how most platforms will access your code. 3. Sign up for a serverless platform: I recommend Google Cloud Run to beginners because its free tier is generous and it's built for AI workloads. 4. Connect and Deploy: Point Cloud Run to your GitHub repository, configure your main file, and hit "Deploy." In a few minutes, you'll have a public URL for your agent. That's it. You've gone from a local script to a live web service. Things That Will Instantly Break in Production Your agent will work differently in the cloud than on your laptop. Here are the traps everyone falls into: * Hardcoded API Keys: If your OpenAI key is sitting in your Python file, you're doing it wrong. All platforms have a "secrets" or "environment variables" section. Put your keys there. This is non negotiable for security. * Forgetting about Memory: Serverless functions are stateless. Your agent won't remember the last conversation unless you connect it to an external database like Redis or a simple cloud SQL instance. * Using Local File Paths: Your script that reads `C:/Users/Dave/Documents/data.csv` will fail immediately. All files need to be accessed from cloud storage (like AWS S3 or Google Cloud Storage) or included in the deployment package itself. Stop trying to build the perfect, infinitely scalable architecture from day one. Get your agent online with the simplest method possible, see how it behaves, and then solve the problems you actually have.
PSA - How to Fix Wrist Pain From Drumming **Hey everyone!** My name is Elliot and I'm a doctor of Physical Therapy who specializes in repetitive strain injury and I have a background working with musicians and the music industry. Over the past few weeks I've noticed an increase in posts about wrist pain, hand fatigue, and discomfort during/after practice sessions. **I wanted to jump in and provide some practical, evidence-based guidance** that I use with the professional drummers and musicians I work with. This isn't just "rest and ice" advice - I want to give you actionable steps to actually prevent and address these issues. **You're not alone, and this isn't just "getting older" or "part of drumming."** These are preventable repetitive strain injuries (RSI) that can be addressed with the right approach. https://preview.redd.it/vu0cw34gfssf1.png?width=840&format=png&auto=webp&s=4c5115f18c2bbf7b5c642c79c20f8ef56fb36f09 Drumming requires: * **Intense, repetitive radial and ulnar deviation** (side-to-side wrist movements) * Thousands of stick impacts per practice session * Sustained grip tension on sticks * Long practice sessions without breaks * High-velocity muscle contractions during fast passages * Constant fine motor control for dynamics # How Your Muscle "Healthbar" Works https://preview.redd.it/akq6dk5kfssf1.png?width=711&format=png&auto=webp&s=1ff40484df44f5bc3af748ad998e6e12cf67c7cb I spent many years working in the gaming industry as well so we built this model of thinking about repetitive strain injuries like a health bar in a video game. **Starting Capacity:** Your initial healthbar size depends on your baseline tissue endurance and conditioning. If you're new to drumming or haven't been physically active, your healthbar starts smaller. **Daily Depletion:** Every stroke gradually depletes your healthbar. The depletion rate depends on: * **Intensity:** How hard you're hitting and gripping * **Duration:** How long you practice without breaks * **Technique:** Inefficient movement patterns waste "HP" * **Tempo:** Faster playing depletes your healthbar quicker * **Stick grip:** Death gripping accelerates depletion **Recovery Threshold:** When your healthbar drops below a critical point, discomfort and pain begin. If you continue practicing without allowing recovery, severe pain and tissue damage develop. # Why Radial & Ulnar Deviation Matter Most in Drumming When you drum, your wrists move primarily in **two planes of motion:** **Radial Deviation:** Wrist moves toward the thumb side (like a hi-hat motion or ride cymbal) **Ulnar Deviation:** Wrist moves toward the pinky side (like coming back from a crash or floor tom) These movements are controlled by specific muscles and tendons that run along the sides of your forearm: https://preview.redd.it/yh436qbxfssf1.png?width=600&format=png&auto=webp&s=6638508347306f3557e4665b4c0fa79e21c68215 **Radial Deviators (Thumb Side):** * Extensor Carpi Radialis Longus & Brevis * Flexor Carpi Radialis * Abductor Pollicis Longus * Extensor Pollicis Brevis **Ulnar Deviators (Pinky Side):** * Extensor Carpi Ulnaris * Flexor Carpi Ulnaris When you play drums, you're performing **thousands of these side-to-side movements** every practice session. A typical rock beat at 120 BPM means 240 strokes per minute - that's **14,400 repetitions in just one hour** of practice! # Areas You Might Feel Pain Drummers most commonly experience pain in these areas: https://preview.redd.it/6ty6xf80gssf1.png?width=936&format=png&auto=webp&s=908721212f0a215957d9c5f51436320adcf8e16d * **Outside of the wrist (ulnar side):** Most common - pain along the pinky side from ulnar deviation overuse * **Top of the forearm (radial side):** Pain along the thumb side near the wrist * **Base of the thumb:** From compensatory gripping patterns * **Inside forearm near the elbow:** From the wrist flexors working overtime * **Outside forearm near the elbow:** Tennis elbow-like symptoms from the extensors The pain typically starts as a dull ache after practice, then progresses to sharp pain during playing if not addressed. # Temporary vs. Permanent Solutions **Temporary:** Rest, ice, and massage restore some immediate HP but don't increase your total capacity **Permanent:** Only consistent endurance training permanently increases your healthbar's maximum capacity # Why Muscle Endurance Matters More Than Strength https://preview.redd.it/jwvvhc8bgssf1.png?width=1273&format=png&auto=webp&s=ad6670ba77a6a85bf5231bec961bfbf485004dce Most drummers don't need to lift heavy weights for drumming - they need muscles that can handle tens of thousands of small, repeated movements without fatigue. This is why: **Endurance Training Benefits:** * Increases your tissue's capacity to handle repetitive stress * Improves blood flow to working muscles * Enhances your nervous system's efficiency in recruiting muscle fibers * Bui
After building 17+ MVPs, here's the tech stack we use for 90% of projects **After building 17+ MVPs, here's the tech stack we use for 90% of projects:** (And why most founders are overthinking it) **The Boring Stack That Makes Money:** * Frontend: Next.js * Backend: Supabase * Payments: Stripe * Hosting: Vercel * Email: Resend * Analytics: PostHog Cost: $86/month Setup time: 2 hours Scales to: Millions **What we DON'T use for MVPs:** ❌ Microservices (you're not Netflix) ❌ Kubernetes (you have 10 users, not 10 million) ❌ Redis (PostgreSQL is fast enough) ❌ React Native (PWA works fine) ❌ TypeScript (controversial: JS for speed, TS for scale) **Real client results with this "basic" stack:** Jake Thompson (3D Omega): 1,000+ businesses 5,000+ artists $32K MRR Still on the $86/month stack Emily Parker: 40% conversion rate 3x faster than previous build 94% customer satisfaction Zero scaling issues **The expensive lesson:** \- We had a client who insisted on "enterprise architecture." \- Microservices. GraphQL. The works. \- 8 months. $100K. Never launched. Their competitor used our boring stack. \- 6 weeks. $3.5K. Now at $50K MRR. **Your tech stack doesn't matter if you have no users.** Instagram scaled to 1 billion on 13 engineers. WhatsApp supported 900 million users with 50 engineers. You're not Instagram. Use boring technology. **The One Rule:** Can you explain your stack in 30 seconds? If not, it's too complex for an MVP. **Our stack explanation:** "Next.js for the app, Supabase for data, Stripe for payments. Done." That's it. That's the stack generating millions in client revenue. Stop debating React vs Vue. Stop optimizing for problems you don't have. Stop architecting for scale you haven't earned. Ship something people pay for. Then worry about the tech. What part of your stack could you delete today?
Why are people leaving Vercel Hosting but not NextJS? I’m curious, why are most people talking about leaving Vercel hosting only and not NextJS as it is owned by Vercel as well? Isn’t that the elephant in the room that we need to address too? For me personally I’m utterly disappointed with what the CEO did starting to transition away from NextJS to other frameworks in addition to Vercel hosting.
It’s not you, it’s Europe I’d rather hear “we are moving forward with another candidate” than the job has moved to Europe 🥲
Built a (french) city builder with Nuxt Hey, I made this browser game where you can build a prehistoric city, it's nothing hard there are no special mechanics but if someone wants to fork the code and add features feel free ;) [https://github.com/TakCastel/prehistopia](https://github.com/TakCastel/prehistopia) Link to the game : [https://prehistopia.vercel.app/](https://prehistopia.vercel.app/) Tell me what you guys think about this project I did in 1 week :)
AW HELL NAH YO ASS TWEAKIN JIGSAW context: rule34dle is a game where you guess which one has more porn
Does anyone else think the whole "separate database provider" trend is completely backwards? Okay so I'm a developer with 15 years of PHP, NodeJS and am studying for Security+ right now and this is driving me crazy. How did we all just... agree that it's totally fine to host your app on one provider and yeet your database onto a completely different one across the public internet? Examples I have found. * Laravel Cloud connecting to some Postgres instance on Neon (possibly the same one according to other posts) * Vercel apps hitting databases on Neon/PlanetScale/Supabase * Upstash Redis The latency is stupid. Every. Single. Query. has to go across the internet now. Yeah yeah, I know about PoPs and edge locations and all that stuff, but you're still adding a massive amount of latency compared to same-VPC or same-datacenter connections. A query that should take like 1-2ms now takes 20-50ms+ because it's doing a round trip through who knows how many networks. And if you've got an N+1 query problem? Your 100ms page just became 5 seconds. And yes, I KNOW it's TLS encrypted. But you're still exposing your database to the entire internet. Your connection strings all of it is traveling across networks you don't own or control. Like I said, I'm studying Security+ right now and I can't even imagine trying to explain to a compliance/security team why customer data is bouncing through the public internet 50 times per page load. That meeting would be... interesting. Look, I get it - the Developer Experience is stupid easy. Click a button, get a connection string, paste it in your env file, deploy. But we're trading actual performance and security for convenience. We're adding latency, more potential failure points, security holes, and locking ourselves into multiple vendors. All so we can skip learning how to properly set up a database? What happened to keeping your database close to your app? VPC peering? Actually caring about performance? What is everyones thoughts on this?
Vercel CEO meets with Netanyahu to discuss AI education
Infinite World in TypeScript # Try it here [https://generate-infinite.vercel.app/](https://generate-infinite.vercel.app/) Been working on an **immersive, infinite, procedurally generated world** built with **Three.js** and **WebGL**, fully refactored in **TypeScript**. This project is a TypeScript-based evolution of the original [Infinite World](https://github.com/brunosimon/infinite-world) concept by the creative developer Bruno Simon. It serves as: * technical demonstration of procedural generation. * a case study in migrating a JavaScript codebase to TypeScript for enhanced scalability and maintainability. # Key Features * **Infinite Procedural Terrain:** Explores a unique, dynamically generated world every time. * **Built with Three.js:** Leverages the power of **Three.js** for efficient **WebGL** rendering. * **Fully Typed Codebase:** Migrated from JavaScript to **TypeScript** for robust, maintainable, and error-free code. * **Enhanced Rendering:** Features improved lighting, shading, and terrain generation algorithms for a more natural look. * **Collision Detection:** Implemented a camera collision system to prevent clipping through the terrain, improving the user experience. # Notes It was tested on Linux, and it works perfectly well, however some rendering issues have been observed when testing on Windows, it is still a mystery, so your insight and feedback are welcome and appreciated! # More details [ibra-kdbra/Generate\_infinite](https://github.com/ibra-kdbra/Generate_infinite)
Everyone Builds AI Agents. Almost No One Knows How to Deploy Them. I've seen this happen a dozen times with clients. A team spends weeks building a brilliant agent with LangChain or CrewAI. It works flawlessly on their laptop. Then they ask the million-dollar question: "So... how do we get this online so people can actually use it?" The silence is deafening. Most tutorials stop right before the most important part. Your agent is a cool science project until it's live. You can't just keep a terminal window open on your machine forever. So here’s the no nonsense guide to actually getting your agent deployed, based on what works in the real world. The Three Places Your Agent Can Actually Live Forget the complex diagrams. For 99% of projects, you have three real options. * Serverless (The "Start Here" Method): This is the default for most new agents. Platforms like Google Cloud Run, Vercel, or even Genezio let you deploy code directly from GitHub without ever thinking about a server. You just provide your code, and they handle the rest. You pay only when the agent is actively running. This is perfect for simple chatbots, Q&A tools, or basic workflow automations. * Containers (The "It's Getting Serious" Method): This is your next step up. You package your agent and all its dependencies into a Docker container. Think of it as a self-contained box that can run anywhere. You then deploy this container to a service like Cloud Run (which also runs containers), AWS ECS, or Azure Container Apps. You do this when your agent needs more memory, has to run for more than a few minutes (like processing a large document), or has finicky dependencies. * Full Servers (The "Don't Do This Yet" Method): This is managing your own virtual machines or using a complex system like Kubernetes. I'm telling you this so you know to avoid it. Unless you're building a massive, enterprise scale platform with thousands of concurrent users, this is a surefire way to waste months on infrastructure instead of improving your agent. A Dead Simple Path for Your First Deployment Don't overthink it. Here is the fastest way to get your first agent live. 1. Wrap your agent in an API: Your Python script needs a way to receive web requests. Use a simple framework like Flask or FastAPI to create a single API endpoint that triggers your agent. 2. Push your code to GitHub: This is standard practice and how most platforms will access your code. 3. Sign up for a serverless platform: I recommend Google Cloud Run to beginners because its free tier is generous and it's built for AI workloads. 4. Connect and Deploy: Point Cloud Run to your GitHub repository, configure your main file, and hit "Deploy." In a few minutes, you'll have a public URL for your agent. That's it. You've gone from a local script to a live web service. Things That Will Instantly Break in Production Your agent will work differently in the cloud than on your laptop. Here are the traps everyone falls into: * Hardcoded API Keys: If your OpenAI key is sitting in your Python file, you're doing it wrong. All platforms have a "secrets" or "environment variables" section. Put your keys there. This is non negotiable for security. * Forgetting about Memory: Serverless functions are stateless. Your agent won't remember the last conversation unless you connect it to an external database like Redis or a simple cloud SQL instance. * Using Local File Paths: Your script that reads `C:/Users/Dave/Documents/data.csv` will fail immediately. All files need to be accessed from cloud storage (like AWS S3 or Google Cloud Storage) or included in the deployment package itself. Stop trying to build the perfect, infinitely scalable architecture from day one. Get your agent online with the simplest method possible, see how it behaves, and then solve the problems you actually have.
PSA - How to Fix Wrist Pain From Drumming **Hey everyone!** My name is Elliot and I'm a doctor of Physical Therapy who specializes in repetitive strain injury and I have a background working with musicians and the music industry. Over the past few weeks I've noticed an increase in posts about wrist pain, hand fatigue, and discomfort during/after practice sessions. **I wanted to jump in and provide some practical, evidence-based guidance** that I use with the professional drummers and musicians I work with. This isn't just "rest and ice" advice - I want to give you actionable steps to actually prevent and address these issues. **You're not alone, and this isn't just "getting older" or "part of drumming."** These are preventable repetitive strain injuries (RSI) that can be addressed with the right approach. https://preview.redd.it/vu0cw34gfssf1.png?width=840&format=png&auto=webp&s=4c5115f18c2bbf7b5c642c79c20f8ef56fb36f09 Drumming requires: * **Intense, repetitive radial and ulnar deviation** (side-to-side wrist movements) * Thousands of stick impacts per practice session * Sustained grip tension on sticks * Long practice sessions without breaks * High-velocity muscle contractions during fast passages * Constant fine motor control for dynamics # How Your Muscle "Healthbar" Works https://preview.redd.it/akq6dk5kfssf1.png?width=711&format=png&auto=webp&s=1ff40484df44f5bc3af748ad998e6e12cf67c7cb I spent many years working in the gaming industry as well so we built this model of thinking about repetitive strain injuries like a health bar in a video game. **Starting Capacity:** Your initial healthbar size depends on your baseline tissue endurance and conditioning. If you're new to drumming or haven't been physically active, your healthbar starts smaller. **Daily Depletion:** Every stroke gradually depletes your healthbar. The depletion rate depends on: * **Intensity:** How hard you're hitting and gripping * **Duration:** How long you practice without breaks * **Technique:** Inefficient movement patterns waste "HP" * **Tempo:** Faster playing depletes your healthbar quicker * **Stick grip:** Death gripping accelerates depletion **Recovery Threshold:** When your healthbar drops below a critical point, discomfort and pain begin. If you continue practicing without allowing recovery, severe pain and tissue damage develop. # Why Radial & Ulnar Deviation Matter Most in Drumming When you drum, your wrists move primarily in **two planes of motion:** **Radial Deviation:** Wrist moves toward the thumb side (like a hi-hat motion or ride cymbal) **Ulnar Deviation:** Wrist moves toward the pinky side (like coming back from a crash or floor tom) These movements are controlled by specific muscles and tendons that run along the sides of your forearm: https://preview.redd.it/yh436qbxfssf1.png?width=600&format=png&auto=webp&s=6638508347306f3557e4665b4c0fa79e21c68215 **Radial Deviators (Thumb Side):** * Extensor Carpi Radialis Longus & Brevis * Flexor Carpi Radialis * Abductor Pollicis Longus * Extensor Pollicis Brevis **Ulnar Deviators (Pinky Side):** * Extensor Carpi Ulnaris * Flexor Carpi Ulnaris When you play drums, you're performing **thousands of these side-to-side movements** every practice session. A typical rock beat at 120 BPM means 240 strokes per minute - that's **14,400 repetitions in just one hour** of practice! # Areas You Might Feel Pain Drummers most commonly experience pain in these areas: https://preview.redd.it/6ty6xf80gssf1.png?width=936&format=png&auto=webp&s=908721212f0a215957d9c5f51436320adcf8e16d * **Outside of the wrist (ulnar side):** Most common - pain along the pinky side from ulnar deviation overuse * **Top of the forearm (radial side):** Pain along the thumb side near the wrist * **Base of the thumb:** From compensatory gripping patterns * **Inside forearm near the elbow:** From the wrist flexors working overtime * **Outside forearm near the elbow:** Tennis elbow-like symptoms from the extensors The pain typically starts as a dull ache after practice, then progresses to sharp pain during playing if not addressed. # Temporary vs. Permanent Solutions **Temporary:** Rest, ice, and massage restore some immediate HP but don't increase your total capacity **Permanent:** Only consistent endurance training permanently increases your healthbar's maximum capacity # Why Muscle Endurance Matters More Than Strength https://preview.redd.it/jwvvhc8bgssf1.png?width=1273&format=png&auto=webp&s=ad6670ba77a6a85bf5231bec961bfbf485004dce Most drummers don't need to lift heavy weights for drumming - they need muscles that can handle tens of thousands of small, repeated movements without fatigue. This is why: **Endurance Training Benefits:** * Increases your tissue's capacity to handle repetitive stress * Improves blood flow to working muscles * Enhances your nervous system's efficiency in recruiting muscle fibers * Bui
After building 17+ MVPs, here's the tech stack we use for 90% of projects **After building 17+ MVPs, here's the tech stack we use for 90% of projects:** (And why most founders are overthinking it) **The Boring Stack That Makes Money:** * Frontend: Next.js * Backend: Supabase * Payments: Stripe * Hosting: Vercel * Email: Resend * Analytics: PostHog Cost: $86/month Setup time: 2 hours Scales to: Millions **What we DON'T use for MVPs:** ❌ Microservices (you're not Netflix) ❌ Kubernetes (you have 10 users, not 10 million) ❌ Redis (PostgreSQL is fast enough) ❌ React Native (PWA works fine) ❌ TypeScript (controversial: JS for speed, TS for scale) **Real client results with this "basic" stack:** Jake Thompson (3D Omega): 1,000+ businesses 5,000+ artists $32K MRR Still on the $86/month stack Emily Parker: 40% conversion rate 3x faster than previous build 94% customer satisfaction Zero scaling issues **The expensive lesson:** \- We had a client who insisted on "enterprise architecture." \- Microservices. GraphQL. The works. \- 8 months. $100K. Never launched. Their competitor used our boring stack. \- 6 weeks. $3.5K. Now at $50K MRR. **Your tech stack doesn't matter if you have no users.** Instagram scaled to 1 billion on 13 engineers. WhatsApp supported 900 million users with 50 engineers. You're not Instagram. Use boring technology. **The One Rule:** Can you explain your stack in 30 seconds? If not, it's too complex for an MVP. **Our stack explanation:** "Next.js for the app, Supabase for data, Stripe for payments. Done." That's it. That's the stack generating millions in client revenue. Stop debating React vs Vue. Stop optimizing for problems you don't have. Stop architecting for scale you haven't earned. Ship something people pay for. Then worry about the tech. What part of your stack could you delete today?
Why are people leaving Vercel Hosting but not NextJS? I’m curious, why are most people talking about leaving Vercel hosting only and not NextJS as it is owned by Vercel as well? Isn’t that the elephant in the room that we need to address too? For me personally I’m utterly disappointed with what the CEO did starting to transition away from NextJS to other frameworks in addition to Vercel hosting.
It’s not you, it’s Europe I’d rather hear “we are moving forward with another candidate” than the job has moved to Europe 🥲
Built a (french) city builder with Nuxt Hey, I made this browser game where you can build a prehistoric city, it's nothing hard there are no special mechanics but if someone wants to fork the code and add features feel free ;) [https://github.com/TakCastel/prehistopia](https://github.com/TakCastel/prehistopia) Link to the game : [https://prehistopia.vercel.app/](https://prehistopia.vercel.app/) Tell me what you guys think about this project I did in 1 week :)
AW HELL NAH YO ASS TWEAKIN JIGSAW context: rule34dle is a game where you guess which one has more porn
Does anyone else think the whole "separate database provider" trend is completely backwards? Okay so I'm a developer with 15 years of PHP, NodeJS and am studying for Security+ right now and this is driving me crazy. How did we all just... agree that it's totally fine to host your app on one provider and yeet your database onto a completely different one across the public internet? Examples I have found. * Laravel Cloud connecting to some Postgres instance on Neon (possibly the same one according to other posts) * Vercel apps hitting databases on Neon/PlanetScale/Supabase * Upstash Redis The latency is stupid. Every. Single. Query. has to go across the internet now. Yeah yeah, I know about PoPs and edge locations and all that stuff, but you're still adding a massive amount of latency compared to same-VPC or same-datacenter connections. A query that should take like 1-2ms now takes 20-50ms+ because it's doing a round trip through who knows how many networks. And if you've got an N+1 query problem? Your 100ms page just became 5 seconds. And yes, I KNOW it's TLS encrypted. But you're still exposing your database to the entire internet. Your connection strings all of it is traveling across networks you don't own or control. Like I said, I'm studying Security+ right now and I can't even imagine trying to explain to a compliance/security team why customer data is bouncing through the public internet 50 times per page load. That meeting would be... interesting. Look, I get it - the Developer Experience is stupid easy. Click a button, get a connection string, paste it in your env file, deploy. But we're trading actual performance and security for convenience. We're adding latency, more potential failure points, security holes, and locking ourselves into multiple vendors. All so we can skip learning how to properly set up a database? What happened to keeping your database close to your app? VPC peering? Actually caring about performance? What is everyones thoughts on this?
Vercel CEO meets with Netanyahu to discuss AI education
Infinite World in TypeScript # Try it here [https://generate-infinite.vercel.app/](https://generate-infinite.vercel.app/) Been working on an **immersive, infinite, procedurally generated world** built with **Three.js** and **WebGL**, fully refactored in **TypeScript**. This project is a TypeScript-based evolution of the original [Infinite World](https://github.com/brunosimon/infinite-world) concept by the creative developer Bruno Simon. It serves as: * technical demonstration of procedural generation. * a case study in migrating a JavaScript codebase to TypeScript for enhanced scalability and maintainability. # Key Features * **Infinite Procedural Terrain:** Explores a unique, dynamically generated world every time. * **Built with Three.js:** Leverages the power of **Three.js** for efficient **WebGL** rendering. * **Fully Typed Codebase:** Migrated from JavaScript to **TypeScript** for robust, maintainable, and error-free code. * **Enhanced Rendering:** Features improved lighting, shading, and terrain generation algorithms for a more natural look. * **Collision Detection:** Implemented a camera collision system to prevent clipping through the terrain, improving the user experience. # Notes It was tested on Linux, and it works perfectly well, however some rendering issues have been observed when testing on Windows, it is still a mystery, so your insight and feedback are welcome and appreciated! # More details [ibra-kdbra/Generate\_infinite](https://github.com/ibra-kdbra/Generate_infinite)
Everyone Builds AI Agents. Almost No One Knows How to Deploy Them. I've seen this happen a dozen times with clients. A team spends weeks building a brilliant agent with LangChain or CrewAI. It works flawlessly on their laptop. Then they ask the million-dollar question: "So... how do we get this online so people can actually use it?" The silence is deafening. Most tutorials stop right before the most important part. Your agent is a cool science project until it's live. You can't just keep a terminal window open on your machine forever. So here’s the no nonsense guide to actually getting your agent deployed, based on what works in the real world. The Three Places Your Agent Can Actually Live Forget the complex diagrams. For 99% of projects, you have three real options. * Serverless (The "Start Here" Method): This is the default for most new agents. Platforms like Google Cloud Run, Vercel, or even Genezio let you deploy code directly from GitHub without ever thinking about a server. You just provide your code, and they handle the rest. You pay only when the agent is actively running. This is perfect for simple chatbots, Q&A tools, or basic workflow automations. * Containers (The "It's Getting Serious" Method): This is your next step up. You package your agent and all its dependencies into a Docker container. Think of it as a self-contained box that can run anywhere. You then deploy this container to a service like Cloud Run (which also runs containers), AWS ECS, or Azure Container Apps. You do this when your agent needs more memory, has to run for more than a few minutes (like processing a large document), or has finicky dependencies. * Full Servers (The "Don't Do This Yet" Method): This is managing your own virtual machines or using a complex system like Kubernetes. I'm telling you this so you know to avoid it. Unless you're building a massive, enterprise scale platform with thousands of concurrent users, this is a surefire way to waste months on infrastructure instead of improving your agent. A Dead Simple Path for Your First Deployment Don't overthink it. Here is the fastest way to get your first agent live. 1. Wrap your agent in an API: Your Python script needs a way to receive web requests. Use a simple framework like Flask or FastAPI to create a single API endpoint that triggers your agent. 2. Push your code to GitHub: This is standard practice and how most platforms will access your code. 3. Sign up for a serverless platform: I recommend Google Cloud Run to beginners because its free tier is generous and it's built for AI workloads. 4. Connect and Deploy: Point Cloud Run to your GitHub repository, configure your main file, and hit "Deploy." In a few minutes, you'll have a public URL for your agent. That's it. You've gone from a local script to a live web service. Things That Will Instantly Break in Production Your agent will work differently in the cloud than on your laptop. Here are the traps everyone falls into: * Hardcoded API Keys: If your OpenAI key is sitting in your Python file, you're doing it wrong. All platforms have a "secrets" or "environment variables" section. Put your keys there. This is non negotiable for security. * Forgetting about Memory: Serverless functions are stateless. Your agent won't remember the last conversation unless you connect it to an external database like Redis or a simple cloud SQL instance. * Using Local File Paths: Your script that reads `C:/Users/Dave/Documents/data.csv` will fail immediately. All files need to be accessed from cloud storage (like AWS S3 or Google Cloud Storage) or included in the deployment package itself. Stop trying to build the perfect, infinitely scalable architecture from day one. Get your agent online with the simplest method possible, see how it behaves, and then solve the problems you actually have.
PSA - How to Fix Wrist Pain From Drumming **Hey everyone!** My name is Elliot and I'm a doctor of Physical Therapy who specializes in repetitive strain injury and I have a background working with musicians and the music industry. Over the past few weeks I've noticed an increase in posts about wrist pain, hand fatigue, and discomfort during/after practice sessions. **I wanted to jump in and provide some practical, evidence-based guidance** that I use with the professional drummers and musicians I work with. This isn't just "rest and ice" advice - I want to give you actionable steps to actually prevent and address these issues. **You're not alone, and this isn't just "getting older" or "part of drumming."** These are preventable repetitive strain injuries (RSI) that can be addressed with the right approach. https://preview.redd.it/vu0cw34gfssf1.png?width=840&format=png&auto=webp&s=4c5115f18c2bbf7b5c642c79c20f8ef56fb36f09 Drumming requires: * **Intense, repetitive radial and ulnar deviation** (side-to-side wrist movements) * Thousands of stick impacts per practice session * Sustained grip tension on sticks * Long practice sessions without breaks * High-velocity muscle contractions during fast passages * Constant fine motor control for dynamics # How Your Muscle "Healthbar" Works https://preview.redd.it/akq6dk5kfssf1.png?width=711&format=png&auto=webp&s=1ff40484df44f5bc3af748ad998e6e12cf67c7cb I spent many years working in the gaming industry as well so we built this model of thinking about repetitive strain injuries like a health bar in a video game. **Starting Capacity:** Your initial healthbar size depends on your baseline tissue endurance and conditioning. If you're new to drumming or haven't been physically active, your healthbar starts smaller. **Daily Depletion:** Every stroke gradually depletes your healthbar. The depletion rate depends on: * **Intensity:** How hard you're hitting and gripping * **Duration:** How long you practice without breaks * **Technique:** Inefficient movement patterns waste "HP" * **Tempo:** Faster playing depletes your healthbar quicker * **Stick grip:** Death gripping accelerates depletion **Recovery Threshold:** When your healthbar drops below a critical point, discomfort and pain begin. If you continue practicing without allowing recovery, severe pain and tissue damage develop. # Why Radial & Ulnar Deviation Matter Most in Drumming When you drum, your wrists move primarily in **two planes of motion:** **Radial Deviation:** Wrist moves toward the thumb side (like a hi-hat motion or ride cymbal) **Ulnar Deviation:** Wrist moves toward the pinky side (like coming back from a crash or floor tom) These movements are controlled by specific muscles and tendons that run along the sides of your forearm: https://preview.redd.it/yh436qbxfssf1.png?width=600&format=png&auto=webp&s=6638508347306f3557e4665b4c0fa79e21c68215 **Radial Deviators (Thumb Side):** * Extensor Carpi Radialis Longus & Brevis * Flexor Carpi Radialis * Abductor Pollicis Longus * Extensor Pollicis Brevis **Ulnar Deviators (Pinky Side):** * Extensor Carpi Ulnaris * Flexor Carpi Ulnaris When you play drums, you're performing **thousands of these side-to-side movements** every practice session. A typical rock beat at 120 BPM means 240 strokes per minute - that's **14,400 repetitions in just one hour** of practice! # Areas You Might Feel Pain Drummers most commonly experience pain in these areas: https://preview.redd.it/6ty6xf80gssf1.png?width=936&format=png&auto=webp&s=908721212f0a215957d9c5f51436320adcf8e16d * **Outside of the wrist (ulnar side):** Most common - pain along the pinky side from ulnar deviation overuse * **Top of the forearm (radial side):** Pain along the thumb side near the wrist * **Base of the thumb:** From compensatory gripping patterns * **Inside forearm near the elbow:** From the wrist flexors working overtime * **Outside forearm near the elbow:** Tennis elbow-like symptoms from the extensors The pain typically starts as a dull ache after practice, then progresses to sharp pain during playing if not addressed. # Temporary vs. Permanent Solutions **Temporary:** Rest, ice, and massage restore some immediate HP but don't increase your total capacity **Permanent:** Only consistent endurance training permanently increases your healthbar's maximum capacity # Why Muscle Endurance Matters More Than Strength https://preview.redd.it/jwvvhc8bgssf1.png?width=1273&format=png&auto=webp&s=ad6670ba77a6a85bf5231bec961bfbf485004dce Most drummers don't need to lift heavy weights for drumming - they need muscles that can handle tens of thousands of small, repeated movements without fatigue. This is why: **Endurance Training Benefits:** * Increases your tissue's capacity to handle repetitive stress * Improves blood flow to working muscles * Enhances your nervous system's efficiency in recruiting muscle fibers * Bui
After building 17+ MVPs, here's the tech stack we use for 90% of projects **After building 17+ MVPs, here's the tech stack we use for 90% of projects:** (And why most founders are overthinking it) **The Boring Stack That Makes Money:** * Frontend: Next.js * Backend: Supabase * Payments: Stripe * Hosting: Vercel * Email: Resend * Analytics: PostHog Cost: $86/month Setup time: 2 hours Scales to: Millions **What we DON'T use for MVPs:** ❌ Microservices (you're not Netflix) ❌ Kubernetes (you have 10 users, not 10 million) ❌ Redis (PostgreSQL is fast enough) ❌ React Native (PWA works fine) ❌ TypeScript (controversial: JS for speed, TS for scale) **Real client results with this "basic" stack:** Jake Thompson (3D Omega): 1,000+ businesses 5,000+ artists $32K MRR Still on the $86/month stack Emily Parker: 40% conversion rate 3x faster than previous build 94% customer satisfaction Zero scaling issues **The expensive lesson:** \- We had a client who insisted on "enterprise architecture." \- Microservices. GraphQL. The works. \- 8 months. $100K. Never launched. Their competitor used our boring stack. \- 6 weeks. $3.5K. Now at $50K MRR. **Your tech stack doesn't matter if you have no users.** Instagram scaled to 1 billion on 13 engineers. WhatsApp supported 900 million users with 50 engineers. You're not Instagram. Use boring technology. **The One Rule:** Can you explain your stack in 30 seconds? If not, it's too complex for an MVP. **Our stack explanation:** "Next.js for the app, Supabase for data, Stripe for payments. Done." That's it. That's the stack generating millions in client revenue. Stop debating React vs Vue. Stop optimizing for problems you don't have. Stop architecting for scale you haven't earned. Ship something people pay for. Then worry about the tech. What part of your stack could you delete today?
Why are people leaving Vercel Hosting but not NextJS? I’m curious, why are most people talking about leaving Vercel hosting only and not NextJS as it is owned by Vercel as well? Isn’t that the elephant in the room that we need to address too? For me personally I’m utterly disappointed with what the CEO did starting to transition away from NextJS to other frameworks in addition to Vercel hosting.
It’s not you, it’s Europe I’d rather hear “we are moving forward with another candidate” than the job has moved to Europe 🥲
Built a (french) city builder with Nuxt Hey, I made this browser game where you can build a prehistoric city, it's nothing hard there are no special mechanics but if someone wants to fork the code and add features feel free ;) [https://github.com/TakCastel/prehistopia](https://github.com/TakCastel/prehistopia) Link to the game : [https://prehistopia.vercel.app/](https://prehistopia.vercel.app/) Tell me what you guys think about this project I did in 1 week :)
AW HELL NAH YO ASS TWEAKIN JIGSAW context: rule34dle is a game where you guess which one has more porn
Does anyone else think the whole "separate database provider" trend is completely backwards? Okay so I'm a developer with 15 years of PHP, NodeJS and am studying for Security+ right now and this is driving me crazy. How did we all just... agree that it's totally fine to host your app on one provider and yeet your database onto a completely different one across the public internet? Examples I have found. * Laravel Cloud connecting to some Postgres instance on Neon (possibly the same one according to other posts) * Vercel apps hitting databases on Neon/PlanetScale/Supabase * Upstash Redis The latency is stupid. Every. Single. Query. has to go across the internet now. Yeah yeah, I know about PoPs and edge locations and all that stuff, but you're still adding a massive amount of latency compared to same-VPC or same-datacenter connections. A query that should take like 1-2ms now takes 20-50ms+ because it's doing a round trip through who knows how many networks. And if you've got an N+1 query problem? Your 100ms page just became 5 seconds. And yes, I KNOW it's TLS encrypted. But you're still exposing your database to the entire internet. Your connection strings all of it is traveling across networks you don't own or control. Like I said, I'm studying Security+ right now and I can't even imagine trying to explain to a compliance/security team why customer data is bouncing through the public internet 50 times per page load. That meeting would be... interesting. Look, I get it - the Developer Experience is stupid easy. Click a button, get a connection string, paste it in your env file, deploy. But we're trading actual performance and security for convenience. We're adding latency, more potential failure points, security holes, and locking ourselves into multiple vendors. All so we can skip learning how to properly set up a database? What happened to keeping your database close to your app? VPC peering? Actually caring about performance? What is everyones thoughts on this?
Vercel CEO meets with Netanyahu to discuss AI education
Infinite World in TypeScript # Try it here [https://generate-infinite.vercel.app/](https://generate-infinite.vercel.app/) Been working on an **immersive, infinite, procedurally generated world** built with **Three.js** and **WebGL**, fully refactored in **TypeScript**. This project is a TypeScript-based evolution of the original [Infinite World](https://github.com/brunosimon/infinite-world) concept by the creative developer Bruno Simon. It serves as: * technical demonstration of procedural generation. * a case study in migrating a JavaScript codebase to TypeScript for enhanced scalability and maintainability. # Key Features * **Infinite Procedural Terrain:** Explores a unique, dynamically generated world every time. * **Built with Three.js:** Leverages the power of **Three.js** for efficient **WebGL** rendering. * **Fully Typed Codebase:** Migrated from JavaScript to **TypeScript** for robust, maintainable, and error-free code. * **Enhanced Rendering:** Features improved lighting, shading, and terrain generation algorithms for a more natural look. * **Collision Detection:** Implemented a camera collision system to prevent clipping through the terrain, improving the user experience. # Notes It was tested on Linux, and it works perfectly well, however some rendering issues have been observed when testing on Windows, it is still a mystery, so your insight and feedback are welcome and appreciated! # More details [ibra-kdbra/Generate\_infinite](https://github.com/ibra-kdbra/Generate_infinite)
Everyone Builds AI Agents. Almost No One Knows How to Deploy Them. I've seen this happen a dozen times with clients. A team spends weeks building a brilliant agent with LangChain or CrewAI. It works flawlessly on their laptop. Then they ask the million-dollar question: "So... how do we get this online so people can actually use it?" The silence is deafening. Most tutorials stop right before the most important part. Your agent is a cool science project until it's live. You can't just keep a terminal window open on your machine forever. So here’s the no nonsense guide to actually getting your agent deployed, based on what works in the real world. The Three Places Your Agent Can Actually Live Forget the complex diagrams. For 99% of projects, you have three real options. * Serverless (The "Start Here" Method): This is the default for most new agents. Platforms like Google Cloud Run, Vercel, or even Genezio let you deploy code directly from GitHub without ever thinking about a server. You just provide your code, and they handle the rest. You pay only when the agent is actively running. This is perfect for simple chatbots, Q&A tools, or basic workflow automations. * Containers (The "It's Getting Serious" Method): This is your next step up. You package your agent and all its dependencies into a Docker container. Think of it as a self-contained box that can run anywhere. You then deploy this container to a service like Cloud Run (which also runs containers), AWS ECS, or Azure Container Apps. You do this when your agent needs more memory, has to run for more than a few minutes (like processing a large document), or has finicky dependencies. * Full Servers (The "Don't Do This Yet" Method): This is managing your own virtual machines or using a complex system like Kubernetes. I'm telling you this so you know to avoid it. Unless you're building a massive, enterprise scale platform with thousands of concurrent users, this is a surefire way to waste months on infrastructure instead of improving your agent. A Dead Simple Path for Your First Deployment Don't overthink it. Here is the fastest way to get your first agent live. 1. Wrap your agent in an API: Your Python script needs a way to receive web requests. Use a simple framework like Flask or FastAPI to create a single API endpoint that triggers your agent. 2. Push your code to GitHub: This is standard practice and how most platforms will access your code. 3. Sign up for a serverless platform: I recommend Google Cloud Run to beginners because its free tier is generous and it's built for AI workloads. 4. Connect and Deploy: Point Cloud Run to your GitHub repository, configure your main file, and hit "Deploy." In a few minutes, you'll have a public URL for your agent. That's it. You've gone from a local script to a live web service. Things That Will Instantly Break in Production Your agent will work differently in the cloud than on your laptop. Here are the traps everyone falls into: * Hardcoded API Keys: If your OpenAI key is sitting in your Python file, you're doing it wrong. All platforms have a "secrets" or "environment variables" section. Put your keys there. This is non negotiable for security. * Forgetting about Memory: Serverless functions are stateless. Your agent won't remember the last conversation unless you connect it to an external database like Redis or a simple cloud SQL instance. * Using Local File Paths: Your script that reads `C:/Users/Dave/Documents/data.csv` will fail immediately. All files need to be accessed from cloud storage (like AWS S3 or Google Cloud Storage) or included in the deployment package itself. Stop trying to build the perfect, infinitely scalable architecture from day one. Get your agent online with the simplest method possible, see how it behaves, and then solve the problems you actually have.
PSA - How to Fix Wrist Pain From Drumming **Hey everyone!** My name is Elliot and I'm a doctor of Physical Therapy who specializes in repetitive strain injury and I have a background working with musicians and the music industry. Over the past few weeks I've noticed an increase in posts about wrist pain, hand fatigue, and discomfort during/after practice sessions. **I wanted to jump in and provide some practical, evidence-based guidance** that I use with the professional drummers and musicians I work with. This isn't just "rest and ice" advice - I want to give you actionable steps to actually prevent and address these issues. **You're not alone, and this isn't just "getting older" or "part of drumming."** These are preventable repetitive strain injuries (RSI) that can be addressed with the right approach. https://preview.redd.it/vu0cw34gfssf1.png?width=840&format=png&auto=webp&s=4c5115f18c2bbf7b5c642c79c20f8ef56fb36f09 Drumming requires: * **Intense, repetitive radial and ulnar deviation** (side-to-side wrist movements) * Thousands of stick impacts per practice session * Sustained grip tension on sticks * Long practice sessions without breaks * High-velocity muscle contractions during fast passages * Constant fine motor control for dynamics # How Your Muscle "Healthbar" Works https://preview.redd.it/akq6dk5kfssf1.png?width=711&format=png&auto=webp&s=1ff40484df44f5bc3af748ad998e6e12cf67c7cb I spent many years working in the gaming industry as well so we built this model of thinking about repetitive strain injuries like a health bar in a video game. **Starting Capacity:** Your initial healthbar size depends on your baseline tissue endurance and conditioning. If you're new to drumming or haven't been physically active, your healthbar starts smaller. **Daily Depletion:** Every stroke gradually depletes your healthbar. The depletion rate depends on: * **Intensity:** How hard you're hitting and gripping * **Duration:** How long you practice without breaks * **Technique:** Inefficient movement patterns waste "HP" * **Tempo:** Faster playing depletes your healthbar quicker * **Stick grip:** Death gripping accelerates depletion **Recovery Threshold:** When your healthbar drops below a critical point, discomfort and pain begin. If you continue practicing without allowing recovery, severe pain and tissue damage develop. # Why Radial & Ulnar Deviation Matter Most in Drumming When you drum, your wrists move primarily in **two planes of motion:** **Radial Deviation:** Wrist moves toward the thumb side (like a hi-hat motion or ride cymbal) **Ulnar Deviation:** Wrist moves toward the pinky side (like coming back from a crash or floor tom) These movements are controlled by specific muscles and tendons that run along the sides of your forearm: https://preview.redd.it/yh436qbxfssf1.png?width=600&format=png&auto=webp&s=6638508347306f3557e4665b4c0fa79e21c68215 **Radial Deviators (Thumb Side):** * Extensor Carpi Radialis Longus & Brevis * Flexor Carpi Radialis * Abductor Pollicis Longus * Extensor Pollicis Brevis **Ulnar Deviators (Pinky Side):** * Extensor Carpi Ulnaris * Flexor Carpi Ulnaris When you play drums, you're performing **thousands of these side-to-side movements** every practice session. A typical rock beat at 120 BPM means 240 strokes per minute - that's **14,400 repetitions in just one hour** of practice! # Areas You Might Feel Pain Drummers most commonly experience pain in these areas: https://preview.redd.it/6ty6xf80gssf1.png?width=936&format=png&auto=webp&s=908721212f0a215957d9c5f51436320adcf8e16d * **Outside of the wrist (ulnar side):** Most common - pain along the pinky side from ulnar deviation overuse * **Top of the forearm (radial side):** Pain along the thumb side near the wrist * **Base of the thumb:** From compensatory gripping patterns * **Inside forearm near the elbow:** From the wrist flexors working overtime * **Outside forearm near the elbow:** Tennis elbow-like symptoms from the extensors The pain typically starts as a dull ache after practice, then progresses to sharp pain during playing if not addressed. # Temporary vs. Permanent Solutions **Temporary:** Rest, ice, and massage restore some immediate HP but don't increase your total capacity **Permanent:** Only consistent endurance training permanently increases your healthbar's maximum capacity # Why Muscle Endurance Matters More Than Strength https://preview.redd.it/jwvvhc8bgssf1.png?width=1273&format=png&auto=webp&s=ad6670ba77a6a85bf5231bec961bfbf485004dce Most drummers don't need to lift heavy weights for drumming - they need muscles that can handle tens of thousands of small, repeated movements without fatigue. This is why: **Endurance Training Benefits:** * Increases your tissue's capacity to handle repetitive stress * Improves blood flow to working muscles * Enhances your nervous system's efficiency in recruiting muscle fibers * Bui
After building 17+ MVPs, here's the tech stack we use for 90% of projects **After building 17+ MVPs, here's the tech stack we use for 90% of projects:** (And why most founders are overthinking it) **The Boring Stack That Makes Money:** * Frontend: Next.js * Backend: Supabase * Payments: Stripe * Hosting: Vercel * Email: Resend * Analytics: PostHog Cost: $86/month Setup time: 2 hours Scales to: Millions **What we DON'T use for MVPs:** ❌ Microservices (you're not Netflix) ❌ Kubernetes (you have 10 users, not 10 million) ❌ Redis (PostgreSQL is fast enough) ❌ React Native (PWA works fine) ❌ TypeScript (controversial: JS for speed, TS for scale) **Real client results with this "basic" stack:** Jake Thompson (3D Omega): 1,000+ businesses 5,000+ artists $32K MRR Still on the $86/month stack Emily Parker: 40% conversion rate 3x faster than previous build 94% customer satisfaction Zero scaling issues **The expensive lesson:** \- We had a client who insisted on "enterprise architecture." \- Microservices. GraphQL. The works. \- 8 months. $100K. Never launched. Their competitor used our boring stack. \- 6 weeks. $3.5K. Now at $50K MRR. **Your tech stack doesn't matter if you have no users.** Instagram scaled to 1 billion on 13 engineers. WhatsApp supported 900 million users with 50 engineers. You're not Instagram. Use boring technology. **The One Rule:** Can you explain your stack in 30 seconds? If not, it's too complex for an MVP. **Our stack explanation:** "Next.js for the app, Supabase for data, Stripe for payments. Done." That's it. That's the stack generating millions in client revenue. Stop debating React vs Vue. Stop optimizing for problems you don't have. Stop architecting for scale you haven't earned. Ship something people pay for. Then worry about the tech. What part of your stack could you delete today?
Why are people leaving Vercel Hosting but not NextJS? I’m curious, why are most people talking about leaving Vercel hosting only and not NextJS as it is owned by Vercel as well? Isn’t that the elephant in the room that we need to address too? For me personally I’m utterly disappointed with what the CEO did starting to transition away from NextJS to other frameworks in addition to Vercel hosting.
It’s not you, it’s Europe I’d rather hear “we are moving forward with another candidate” than the job has moved to Europe 🥲
Built a (french) city builder with Nuxt Hey, I made this browser game where you can build a prehistoric city, it's nothing hard there are no special mechanics but if someone wants to fork the code and add features feel free ;) [https://github.com/TakCastel/prehistopia](https://github.com/TakCastel/prehistopia) Link to the game : [https://prehistopia.vercel.app/](https://prehistopia.vercel.app/) Tell me what you guys think about this project I did in 1 week :)
Pick your Vercel alternative only after weighing the pros and cons This sub has had many posts suggesting Vercel alternatives in the last few days. While some suggestions have been solid, others have been outright wrong. IMO it is super-vital to think through each alternative's benefits and limitations before choosing since hosting can get complicated to migrate. * Netlify - DDoS protection and WAF aren't included in non-enterprise plan. On a serverless offering, this can cause billing shocks. * Cloudflare - Nice for SSG and CDN pricing is awesome. But for SSR - Cloudflare Workers run on V8 runtime (and not Node) so every library that works on Node may not readily work. * Self-hosted VPS with Coolify (my preferred choice) - Best budget-wise, no platform locking, but needs initial build & deploy setup. * Railway - Nice predictable pricing, good build & deploy DX, doesn't offer CDN so need to combine with something like Cloudflare. * AWS / GCP services - Make good sense if you are already using these cloud providers, otherwise overwhelming number of offerings and options. Choose wisely, fellas!
Built a web app to help tracking currency spent and share your crafted items. Hello there, I build a web app to track the currency you spent on crafting an item, and at the end it will give you the option to share your item after pasting an image of the crafted item. which will give you the image in the description. \- You can add some currencies to be your favorite, so that u dont have to switch tabs every now and then. Its a very simple app that I can use to share my crafts with my friends, feel free to use it as well <3 Looking for your feedback to make it better. [https://poe-2-crafting-tracker.vercel.app/](https://poe-2-crafting-tracker.vercel.app/)
Vercel Controversy: Ethics, Backlash, and a Migration Guide to Netlify
[Resource] Attendance Calculator | Fully automatic https://preview.redd.it/38f8qocjr1tf1.png?width=1160&format=png&auto=webp&s=c98d369dfdbd4b0205f3aa3d61e5b8e10c645b34 https://preview.redd.it/bb1qmx0lr1tf1.png?width=1142&format=png&auto=webp&s=395b61c481a2a01ff5b9057c05451dbc5aecfd86 https://preview.redd.it/g0sr5aumr1tf1.png?width=1207&format=png&auto=webp&s=cfaad1ad929ee54e7f3482ccb5927a132ba169c2 https://preview.redd.it/biaivw2pr1tf1.png?width=1128&format=png&auto=webp&s=97626a40c7f70d3d0d659dd8699f777792417f1c Website: [Click](https://pesuattendance.vercel.app/) Features: * Connects to your PESU Academy account. * Shows your current attendance for each subject. * Live-calculates how many classes you can miss with a slider. * Works for both RR / EC Campus * Automatically loads calender and holidays * User/Pass are not saved on the server ensuring no logging * Smooth Fluid UI **Appreciate feedback / suggestions** Will clean up the code and put it on GitHub soon. Hope it's useful.
A Civil Rights and ICE tracking tool. Resharing in light of all the increasing fascism The insanity of everything going on with this whole ICE gestapo bullshit is mind boggling. And with growing concerns for our civil liberties, having access to safe, reliable tools is more important than ever. I built Firewatch to empower and protect our communities, and I want to share it again for anyone who feels at risk. Especially since they are taking down other apps from mobile app stores. It is a free and open-source app providing real-time ICE raid alerts, a nationwide map of community-reported incidents, secure incident reporting, Know Your Rights education in multiple languages, legal-grade evidence recording, emergency response tools, and sanctuary resources—all fully private, locally stored, and accessible offline. Just add your own free Gemini API key. If you don't want to use the AI features (the main AI feature is the ability to search your local area for incident and community reports), no worries! 90% of the app will still function perfectly without it. No catch, completely legal, and protected by the 1st Amendment (well...at least it used to be 🙃). Available now to protect civil rights and keep communities safe: https://firewatch-ice.vercel.app This app is built on a foundation of total privacy: It is serverless. There is no central server that I or anyone else can access. It does not track you or store your data. All information is stored locally on your device and is under your control. It is 100% open-source. Every line of code is public on GitHub for anyone to inspect and verify: https://github.com/severian42/Firewatch Key Features: - Real-Time ICE Raid Alerts: Stay informed with community-reported alerts. - Know Your Rights: Essential legal information available in multiple languages, accessible offline. - Secure Incident Reporting & Evidence Recording: Document events safely with legal-grade evidence tools. - Emergency Response & Sanctuary Resources: Quickly access help and find safe locations. Stay safe, know your rights, and support your community. App: https://firewatch-ice.vercel.app Code: https://github.com/severian42/Firewatch Side note: Also, if anyone has any other ideas for apps that could be built as tools to fight back and help protect each other, please drop them in the comments. I want to build more tools for the communities and people who need it. It seems that we, as the civilian population, are the only ones willing to do anything about this administrations insanity on the ground level
Best 8 Free AI Image Generators shared this week Thanks to everyone that posted a free AI image generator this week. I have tested all the links that were shared this week about Free AI Image generators. It is amazing that so many of them exist and with high quality and fast generations. Therefore I put together the list of best ai image generators found on this subreddit this week below. It will save you a lot of money and the time you spend searching for Free AI image generator is greatly reduced. Please let me know if I missed anyone mentioned this week. or comment below **The list can be found below:** 1. [https://aifreeforever.com/image-generators](https://aifreeforever.com/image-generators) \- Top quality free AI images with good prompting. offers unlimited AI Image generation per day without signing up nor login in. Contains some ads. Contains 120 seconds time before next generation. Quality is amazing. slightly can process adult nsfw prompts but converts it to bikinni clothing. Text generation is average. 2. [https://freegen.app/](https://freegen.app/) \- Completely free with no signup required. Instant access to AI art creation. Simple and straightforward interface 3. [https://eyedance.net/es](https://eyedance.net/es) \- Truly unlimited free generations with zero login requirements. Spanish interface but works internationally. Uses Flux Krea and Flux Kontext models. Zero data retention - complete privacy. Takes 10-30 seconds per image. Supports multiple formats and sizes up to 1024x1024. Has image editing features too. Works great on mobile. You own all generated images. 4. [https://pollinations.ai/](https://pollinations.ai/) \- Free open-source API platform. Requires JavaScript enabled. No local hardware needed - everything's cloud-based. 5. [https://visualgpt.io/ai-models/nano-banana](https://visualgpt.io/ai-models/nano-banana) \- Most likely free tier and definitely not unlimited but quality is high. you can generate without signup. i dont know their limit. you can comment when you test. Really fast and consistent results. Great for backgrounds, details, and portrait enhancements. Saves hours compared to Photoshop. Easy for non-technical users. Works within VisualGPT's larger tool ecosystem. 6. [https://dreamator-ai.vercel.app/](https://dreamator-ai.vercel.app/) \- Hosted on Vercel platform - means it's fast and modern. Clean minimalist interface. Likely offers free tier or demo access. Global CDN makes it quick worldwide. Focuses on simple text-to-image generation. Reliable uptime and automatic scaling. 7. [https://www.createimg.com/](https://www.createimg.com/) \- In addition to image generation, it has AI Background Changer for swapping backgrounds. AI Cloth Changer lets you try different outfits. AI Hairstyle to test different hair looks. Colorize Photo for old black and white images. AI Girl for character generation. AI Cartoonizer turns photos into cartoon style. More editing-focused than creation-focused. 8. [https://stable-diffusion-web.com/models](https://stable-diffusion-web.com/models) \- SD 3 Medium for faster experiments. SD 3.5 Large for high-end professional work. Great for learning which model fits your needs. Helps understand speed vs quality tradeoffs. Let me know if you want me to review test and add your image generator to the list. If you have tested and have coments on any one of them please comment below.
Developers drop Vercel, call for boycott after CEO posts selfie with Netanyahu
Codex so far I just upgraded to the Pro plan recently. This is unrelated but ChatGPT Pro with MCPs really feel like it’s able to give novel ideas and find breakthrough research. Anyway, I’ve been coding day and night with Codex and limits are nowhere to be seen which is great. But holy sh*t MCPs with Codex are just absolutely NUTS. I’ve been using different MCPs from Smithery and it’s been really useful so far. But this is where it gets me - I was thinking - hm my Vercel build isn’t working for this new project it’s kinda frustrating. It’s not a mini project I wanna give this much time to for something so simple of an issue. I spent a few mins and I was like you know what? Why if I let codex figure this out. Minute 1 -> Find Vercel MCP Minute 2 -> Add it to Codex very easily Minute ~5 -> Codex is like here’s what I found after scouring through the entire project in Vercel (build logs, deployments, etc), go change this setting And it worked!! Absolutely flawless. What I’m trying to say is, the ‘method’ of doing things is so much more efficient now. *As long as you have the security in mind* (like I excluded the deploy to Vercel function in the mcp), you can get stuff done 500x better than your competitor who doesn’t want to/know how to leverage these. Thank you for everyone who made Codex possible.
Built a little browser game using NOTHING BUT AI tools You can check it out here: [https://the-best-game-in-the-world.vercel.app/](https://the-best-game-in-the-world.vercel.app/) Took just a couple hours of very interrupted work. I used Claude Opus 4.1 and it basically did all of the code. The 3d assets (car and house) were text to 3d model generated using 3daistudio... you could probably prompt out better models than me honestly. Done using threejs, you can give it a try on mobile or desktop, both work. What do you guys think? Is anyone here working on anything remotely similar?
ELDENDLE FIRST UPDATE! Hey guys! 👋 Thank you all for the feedback on the first post. I’ve read every comment and suggestion, and I’m already working on bigger updates based on what you shared. This first patch is a small one, focused on community requests and quality-of-life improvements to make the experience smoother. **✨ What’s new:** * Added **orange feedback for partial matches!** * Improved the win modal, **with a button to switch directly between challenges.** * Added attempt counter for unlocking hints. * Added a timer for the next challenge at the top of the splash page. * Fixed duplicate input bug (pressing Enter with empty input). * Added new Boss Challenge category: **Story Relevance.** (Story Relevance indicates how important a boss is to the main storyline. **“Mandatory”** bosses are required to progress, while **“Optional”** ones are side encounters.) * Improved data accuracy for multiple bosses. * General polish and stability improvements. * Added **7 new characters**: * Commander Niall * Ancestor Spirit * Tibia Mariner * White Mask Varré * Melina * Miriel, Pastor of Vows * Smithing Master Hewg More updates are on the way soon, including new modes and features inspired by your feedback! Thanks again for playing and supporting Eldendle 💛 Play here: [eldendle.vercel.app](http://eldendle.vercel.app/) https://preview.redd.it/2e60lsuhj4tf1.png?width=1899&format=png&auto=webp&s=fe847199a83d7bac0843eb9a74736be8cb89fd70 https://preview.redd.it/spvxj2cjj4tf1.png?width=1299&format=png&auto=webp&s=7d2ae480fa553355057a00d76e3ff2df7658f9f1
Pick your Vercel alternative only after weighing the pros and cons This sub has had many posts suggesting Vercel alternatives in the last few days. While some suggestions have been solid, others have been outright wrong. IMO it is super-vital to think through each alternative's benefits and limitations before choosing since hosting can get complicated to migrate. * Netlify - DDoS protection and WAF aren't included in non-enterprise plan. On a serverless offering, this can cause billing shocks. * Cloudflare - Nice for SSG and CDN pricing is awesome. But for SSR - Cloudflare Workers run on V8 runtime (and not Node) so every library that works on Node may not readily work. * Self-hosted VPS with Coolify (my preferred choice) - Best budget-wise, no platform locking, but needs initial build & deploy setup. * Railway - Nice predictable pricing, good build & deploy DX, doesn't offer CDN so need to combine with something like Cloudflare. * AWS / GCP services - Make good sense if you are already using these cloud providers, otherwise overwhelming number of offerings and options. Choose wisely, fellas!
Built a web app to help tracking currency spent and share your crafted items. Hello there, I build a web app to track the currency you spent on crafting an item, and at the end it will give you the option to share your item after pasting an image of the crafted item. which will give you the image in the description. \- You can add some currencies to be your favorite, so that u dont have to switch tabs every now and then. Its a very simple app that I can use to share my crafts with my friends, feel free to use it as well <3 Looking for your feedback to make it better. [https://poe-2-crafting-tracker.vercel.app/](https://poe-2-crafting-tracker.vercel.app/)
Vercel Controversy: Ethics, Backlash, and a Migration Guide to Netlify
[Resource] Attendance Calculator | Fully automatic https://preview.redd.it/38f8qocjr1tf1.png?width=1160&format=png&auto=webp&s=c98d369dfdbd4b0205f3aa3d61e5b8e10c645b34 https://preview.redd.it/bb1qmx0lr1tf1.png?width=1142&format=png&auto=webp&s=395b61c481a2a01ff5b9057c05451dbc5aecfd86 https://preview.redd.it/g0sr5aumr1tf1.png?width=1207&format=png&auto=webp&s=cfaad1ad929ee54e7f3482ccb5927a132ba169c2 https://preview.redd.it/biaivw2pr1tf1.png?width=1128&format=png&auto=webp&s=97626a40c7f70d3d0d659dd8699f777792417f1c Website: [Click](https://pesuattendance.vercel.app/) Features: * Connects to your PESU Academy account. * Shows your current attendance for each subject. * Live-calculates how many classes you can miss with a slider. * Works for both RR / EC Campus * Automatically loads calender and holidays * User/Pass are not saved on the server ensuring no logging * Smooth Fluid UI **Appreciate feedback / suggestions** Will clean up the code and put it on GitHub soon. Hope it's useful.
A Civil Rights and ICE tracking tool. Resharing in light of all the increasing fascism The insanity of everything going on with this whole ICE gestapo bullshit is mind boggling. And with growing concerns for our civil liberties, having access to safe, reliable tools is more important than ever. I built Firewatch to empower and protect our communities, and I want to share it again for anyone who feels at risk. Especially since they are taking down other apps from mobile app stores. It is a free and open-source app providing real-time ICE raid alerts, a nationwide map of community-reported incidents, secure incident reporting, Know Your Rights education in multiple languages, legal-grade evidence recording, emergency response tools, and sanctuary resources—all fully private, locally stored, and accessible offline. Just add your own free Gemini API key. If you don't want to use the AI features (the main AI feature is the ability to search your local area for incident and community reports), no worries! 90% of the app will still function perfectly without it. No catch, completely legal, and protected by the 1st Amendment (well...at least it used to be 🙃). Available now to protect civil rights and keep communities safe: https://firewatch-ice.vercel.app This app is built on a foundation of total privacy: It is serverless. There is no central server that I or anyone else can access. It does not track you or store your data. All information is stored locally on your device and is under your control. It is 100% open-source. Every line of code is public on GitHub for anyone to inspect and verify: https://github.com/severian42/Firewatch Key Features: - Real-Time ICE Raid Alerts: Stay informed with community-reported alerts. - Know Your Rights: Essential legal information available in multiple languages, accessible offline. - Secure Incident Reporting & Evidence Recording: Document events safely with legal-grade evidence tools. - Emergency Response & Sanctuary Resources: Quickly access help and find safe locations. Stay safe, know your rights, and support your community. App: https://firewatch-ice.vercel.app Code: https://github.com/severian42/Firewatch Side note: Also, if anyone has any other ideas for apps that could be built as tools to fight back and help protect each other, please drop them in the comments. I want to build more tools for the communities and people who need it. It seems that we, as the civilian population, are the only ones willing to do anything about this administrations insanity on the ground level
Best 8 Free AI Image Generators shared this week Thanks to everyone that posted a free AI image generator this week. I have tested all the links that were shared this week about Free AI Image generators. It is amazing that so many of them exist and with high quality and fast generations. Therefore I put together the list of best ai image generators found on this subreddit this week below. It will save you a lot of money and the time you spend searching for Free AI image generator is greatly reduced. Please let me know if I missed anyone mentioned this week. or comment below **The list can be found below:** 1. [https://aifreeforever.com/image-generators](https://aifreeforever.com/image-generators) \- Top quality free AI images with good prompting. offers unlimited AI Image generation per day without signing up nor login in. Contains some ads. Contains 120 seconds time before next generation. Quality is amazing. slightly can process adult nsfw prompts but converts it to bikinni clothing. Text generation is average. 2. [https://freegen.app/](https://freegen.app/) \- Completely free with no signup required. Instant access to AI art creation. Simple and straightforward interface 3. [https://eyedance.net/es](https://eyedance.net/es) \- Truly unlimited free generations with zero login requirements. Spanish interface but works internationally. Uses Flux Krea and Flux Kontext models. Zero data retention - complete privacy. Takes 10-30 seconds per image. Supports multiple formats and sizes up to 1024x1024. Has image editing features too. Works great on mobile. You own all generated images. 4. [https://pollinations.ai/](https://pollinations.ai/) \- Free open-source API platform. Requires JavaScript enabled. No local hardware needed - everything's cloud-based. 5. [https://visualgpt.io/ai-models/nano-banana](https://visualgpt.io/ai-models/nano-banana) \- Most likely free tier and definitely not unlimited but quality is high. you can generate without signup. i dont know their limit. you can comment when you test. Really fast and consistent results. Great for backgrounds, details, and portrait enhancements. Saves hours compared to Photoshop. Easy for non-technical users. Works within VisualGPT's larger tool ecosystem. 6. [https://dreamator-ai.vercel.app/](https://dreamator-ai.vercel.app/) \- Hosted on Vercel platform - means it's fast and modern. Clean minimalist interface. Likely offers free tier or demo access. Global CDN makes it quick worldwide. Focuses on simple text-to-image generation. Reliable uptime and automatic scaling. 7. [https://www.createimg.com/](https://www.createimg.com/) \- In addition to image generation, it has AI Background Changer for swapping backgrounds. AI Cloth Changer lets you try different outfits. AI Hairstyle to test different hair looks. Colorize Photo for old black and white images. AI Girl for character generation. AI Cartoonizer turns photos into cartoon style. More editing-focused than creation-focused. 8. [https://stable-diffusion-web.com/models](https://stable-diffusion-web.com/models) \- SD 3 Medium for faster experiments. SD 3.5 Large for high-end professional work. Great for learning which model fits your needs. Helps understand speed vs quality tradeoffs. Let me know if you want me to review test and add your image generator to the list. If you have tested and have coments on any one of them please comment below.
Developers drop Vercel, call for boycott after CEO posts selfie with Netanyahu
Codex so far I just upgraded to the Pro plan recently. This is unrelated but ChatGPT Pro with MCPs really feel like it’s able to give novel ideas and find breakthrough research. Anyway, I’ve been coding day and night with Codex and limits are nowhere to be seen which is great. But holy sh*t MCPs with Codex are just absolutely NUTS. I’ve been using different MCPs from Smithery and it’s been really useful so far. But this is where it gets me - I was thinking - hm my Vercel build isn’t working for this new project it’s kinda frustrating. It’s not a mini project I wanna give this much time to for something so simple of an issue. I spent a few mins and I was like you know what? Why if I let codex figure this out. Minute 1 -> Find Vercel MCP Minute 2 -> Add it to Codex very easily Minute ~5 -> Codex is like here’s what I found after scouring through the entire project in Vercel (build logs, deployments, etc), go change this setting And it worked!! Absolutely flawless. What I’m trying to say is, the ‘method’ of doing things is so much more efficient now. *As long as you have the security in mind* (like I excluded the deploy to Vercel function in the mcp), you can get stuff done 500x better than your competitor who doesn’t want to/know how to leverage these. Thank you for everyone who made Codex possible.
Built a little browser game using NOTHING BUT AI tools You can check it out here: [https://the-best-game-in-the-world.vercel.app/](https://the-best-game-in-the-world.vercel.app/) Took just a couple hours of very interrupted work. I used Claude Opus 4.1 and it basically did all of the code. The 3d assets (car and house) were text to 3d model generated using 3daistudio... you could probably prompt out better models than me honestly. Done using threejs, you can give it a try on mobile or desktop, both work. What do you guys think? Is anyone here working on anything remotely similar?
ELDENDLE FIRST UPDATE! Hey guys! 👋 Thank you all for the feedback on the first post. I’ve read every comment and suggestion, and I’m already working on bigger updates based on what you shared. This first patch is a small one, focused on community requests and quality-of-life improvements to make the experience smoother. **✨ What’s new:** * Added **orange feedback for partial matches!** * Improved the win modal, **with a button to switch directly between challenges.** * Added attempt counter for unlocking hints. * Added a timer for the next challenge at the top of the splash page. * Fixed duplicate input bug (pressing Enter with empty input). * Added new Boss Challenge category: **Story Relevance.** (Story Relevance indicates how important a boss is to the main storyline. **“Mandatory”** bosses are required to progress, while **“Optional”** ones are side encounters.) * Improved data accuracy for multiple bosses. * General polish and stability improvements. * Added **7 new characters**: * Commander Niall * Ancestor Spirit * Tibia Mariner * White Mask Varré * Melina * Miriel, Pastor of Vows * Smithing Master Hewg More updates are on the way soon, including new modes and features inspired by your feedback! Thanks again for playing and supporting Eldendle 💛 Play here: [eldendle.vercel.app](http://eldendle.vercel.app/) https://preview.redd.it/2e60lsuhj4tf1.png?width=1899&format=png&auto=webp&s=fe847199a83d7bac0843eb9a74736be8cb89fd70 https://preview.redd.it/spvxj2cjj4tf1.png?width=1299&format=png&auto=webp&s=7d2ae480fa553355057a00d76e3ff2df7658f9f1
Pick your Vercel alternative only after weighing the pros and cons This sub has had many posts suggesting Vercel alternatives in the last few days. While some suggestions have been solid, others have been outright wrong. IMO it is super-vital to think through each alternative's benefits and limitations before choosing since hosting can get complicated to migrate. * Netlify - DDoS protection and WAF aren't included in non-enterprise plan. On a serverless offering, this can cause billing shocks. * Cloudflare - Nice for SSG and CDN pricing is awesome. But for SSR - Cloudflare Workers run on V8 runtime (and not Node) so every library that works on Node may not readily work. * Self-hosted VPS with Coolify (my preferred choice) - Best budget-wise, no platform locking, but needs initial build & deploy setup. * Railway - Nice predictable pricing, good build & deploy DX, doesn't offer CDN so need to combine with something like Cloudflare. * AWS / GCP services - Make good sense if you are already using these cloud providers, otherwise overwhelming number of offerings and options. Choose wisely, fellas!
Built a web app to help tracking currency spent and share your crafted items. Hello there, I build a web app to track the currency you spent on crafting an item, and at the end it will give you the option to share your item after pasting an image of the crafted item. which will give you the image in the description. \- You can add some currencies to be your favorite, so that u dont have to switch tabs every now and then. Its a very simple app that I can use to share my crafts with my friends, feel free to use it as well <3 Looking for your feedback to make it better. [https://poe-2-crafting-tracker.vercel.app/](https://poe-2-crafting-tracker.vercel.app/)
Vercel Controversy: Ethics, Backlash, and a Migration Guide to Netlify
[Resource] Attendance Calculator | Fully automatic https://preview.redd.it/38f8qocjr1tf1.png?width=1160&format=png&auto=webp&s=c98d369dfdbd4b0205f3aa3d61e5b8e10c645b34 https://preview.redd.it/bb1qmx0lr1tf1.png?width=1142&format=png&auto=webp&s=395b61c481a2a01ff5b9057c05451dbc5aecfd86 https://preview.redd.it/g0sr5aumr1tf1.png?width=1207&format=png&auto=webp&s=cfaad1ad929ee54e7f3482ccb5927a132ba169c2 https://preview.redd.it/biaivw2pr1tf1.png?width=1128&format=png&auto=webp&s=97626a40c7f70d3d0d659dd8699f777792417f1c Website: [Click](https://pesuattendance.vercel.app/) Features: * Connects to your PESU Academy account. * Shows your current attendance for each subject. * Live-calculates how many classes you can miss with a slider. * Works for both RR / EC Campus * Automatically loads calender and holidays * User/Pass are not saved on the server ensuring no logging * Smooth Fluid UI **Appreciate feedback / suggestions** Will clean up the code and put it on GitHub soon. Hope it's useful.
A Civil Rights and ICE tracking tool. Resharing in light of all the increasing fascism The insanity of everything going on with this whole ICE gestapo bullshit is mind boggling. And with growing concerns for our civil liberties, having access to safe, reliable tools is more important than ever. I built Firewatch to empower and protect our communities, and I want to share it again for anyone who feels at risk. Especially since they are taking down other apps from mobile app stores. It is a free and open-source app providing real-time ICE raid alerts, a nationwide map of community-reported incidents, secure incident reporting, Know Your Rights education in multiple languages, legal-grade evidence recording, emergency response tools, and sanctuary resources—all fully private, locally stored, and accessible offline. Just add your own free Gemini API key. If you don't want to use the AI features (the main AI feature is the ability to search your local area for incident and community reports), no worries! 90% of the app will still function perfectly without it. No catch, completely legal, and protected by the 1st Amendment (well...at least it used to be 🙃). Available now to protect civil rights and keep communities safe: https://firewatch-ice.vercel.app This app is built on a foundation of total privacy: It is serverless. There is no central server that I or anyone else can access. It does not track you or store your data. All information is stored locally on your device and is under your control. It is 100% open-source. Every line of code is public on GitHub for anyone to inspect and verify: https://github.com/severian42/Firewatch Key Features: - Real-Time ICE Raid Alerts: Stay informed with community-reported alerts. - Know Your Rights: Essential legal information available in multiple languages, accessible offline. - Secure Incident Reporting & Evidence Recording: Document events safely with legal-grade evidence tools. - Emergency Response & Sanctuary Resources: Quickly access help and find safe locations. Stay safe, know your rights, and support your community. App: https://firewatch-ice.vercel.app Code: https://github.com/severian42/Firewatch Side note: Also, if anyone has any other ideas for apps that could be built as tools to fight back and help protect each other, please drop them in the comments. I want to build more tools for the communities and people who need it. It seems that we, as the civilian population, are the only ones willing to do anything about this administrations insanity on the ground level
Best 8 Free AI Image Generators shared this week Thanks to everyone that posted a free AI image generator this week. I have tested all the links that were shared this week about Free AI Image generators. It is amazing that so many of them exist and with high quality and fast generations. Therefore I put together the list of best ai image generators found on this subreddit this week below. It will save you a lot of money and the time you spend searching for Free AI image generator is greatly reduced. Please let me know if I missed anyone mentioned this week. or comment below **The list can be found below:** 1. [https://aifreeforever.com/image-generators](https://aifreeforever.com/image-generators) \- Top quality free AI images with good prompting. offers unlimited AI Image generation per day without signing up nor login in. Contains some ads. Contains 120 seconds time before next generation. Quality is amazing. slightly can process adult nsfw prompts but converts it to bikinni clothing. Text generation is average. 2. [https://freegen.app/](https://freegen.app/) \- Completely free with no signup required. Instant access to AI art creation. Simple and straightforward interface 3. [https://eyedance.net/es](https://eyedance.net/es) \- Truly unlimited free generations with zero login requirements. Spanish interface but works internationally. Uses Flux Krea and Flux Kontext models. Zero data retention - complete privacy. Takes 10-30 seconds per image. Supports multiple formats and sizes up to 1024x1024. Has image editing features too. Works great on mobile. You own all generated images. 4. [https://pollinations.ai/](https://pollinations.ai/) \- Free open-source API platform. Requires JavaScript enabled. No local hardware needed - everything's cloud-based. 5. [https://visualgpt.io/ai-models/nano-banana](https://visualgpt.io/ai-models/nano-banana) \- Most likely free tier and definitely not unlimited but quality is high. you can generate without signup. i dont know their limit. you can comment when you test. Really fast and consistent results. Great for backgrounds, details, and portrait enhancements. Saves hours compared to Photoshop. Easy for non-technical users. Works within VisualGPT's larger tool ecosystem. 6. [https://dreamator-ai.vercel.app/](https://dreamator-ai.vercel.app/) \- Hosted on Vercel platform - means it's fast and modern. Clean minimalist interface. Likely offers free tier or demo access. Global CDN makes it quick worldwide. Focuses on simple text-to-image generation. Reliable uptime and automatic scaling. 7. [https://www.createimg.com/](https://www.createimg.com/) \- In addition to image generation, it has AI Background Changer for swapping backgrounds. AI Cloth Changer lets you try different outfits. AI Hairstyle to test different hair looks. Colorize Photo for old black and white images. AI Girl for character generation. AI Cartoonizer turns photos into cartoon style. More editing-focused than creation-focused. 8. [https://stable-diffusion-web.com/models](https://stable-diffusion-web.com/models) \- SD 3 Medium for faster experiments. SD 3.5 Large for high-end professional work. Great for learning which model fits your needs. Helps understand speed vs quality tradeoffs. Let me know if you want me to review test and add your image generator to the list. If you have tested and have coments on any one of them please comment below.
Developers drop Vercel, call for boycott after CEO posts selfie with Netanyahu
Codex so far I just upgraded to the Pro plan recently. This is unrelated but ChatGPT Pro with MCPs really feel like it’s able to give novel ideas and find breakthrough research. Anyway, I’ve been coding day and night with Codex and limits are nowhere to be seen which is great. But holy sh*t MCPs with Codex are just absolutely NUTS. I’ve been using different MCPs from Smithery and it’s been really useful so far. But this is where it gets me - I was thinking - hm my Vercel build isn’t working for this new project it’s kinda frustrating. It’s not a mini project I wanna give this much time to for something so simple of an issue. I spent a few mins and I was like you know what? Why if I let codex figure this out. Minute 1 -> Find Vercel MCP Minute 2 -> Add it to Codex very easily Minute ~5 -> Codex is like here’s what I found after scouring through the entire project in Vercel (build logs, deployments, etc), go change this setting And it worked!! Absolutely flawless. What I’m trying to say is, the ‘method’ of doing things is so much more efficient now. *As long as you have the security in mind* (like I excluded the deploy to Vercel function in the mcp), you can get stuff done 500x better than your competitor who doesn’t want to/know how to leverage these. Thank you for everyone who made Codex possible.
Built a little browser game using NOTHING BUT AI tools You can check it out here: [https://the-best-game-in-the-world.vercel.app/](https://the-best-game-in-the-world.vercel.app/) Took just a couple hours of very interrupted work. I used Claude Opus 4.1 and it basically did all of the code. The 3d assets (car and house) were text to 3d model generated using 3daistudio... you could probably prompt out better models than me honestly. Done using threejs, you can give it a try on mobile or desktop, both work. What do you guys think? Is anyone here working on anything remotely similar?
ELDENDLE FIRST UPDATE! Hey guys! 👋 Thank you all for the feedback on the first post. I’ve read every comment and suggestion, and I’m already working on bigger updates based on what you shared. This first patch is a small one, focused on community requests and quality-of-life improvements to make the experience smoother. **✨ What’s new:** * Added **orange feedback for partial matches!** * Improved the win modal, **with a button to switch directly between challenges.** * Added attempt counter for unlocking hints. * Added a timer for the next challenge at the top of the splash page. * Fixed duplicate input bug (pressing Enter with empty input). * Added new Boss Challenge category: **Story Relevance.** (Story Relevance indicates how important a boss is to the main storyline. **“Mandatory”** bosses are required to progress, while **“Optional”** ones are side encounters.) * Improved data accuracy for multiple bosses. * General polish and stability improvements. * Added **7 new characters**: * Commander Niall * Ancestor Spirit * Tibia Mariner * White Mask Varré * Melina * Miriel, Pastor of Vows * Smithing Master Hewg More updates are on the way soon, including new modes and features inspired by your feedback! Thanks again for playing and supporting Eldendle 💛 Play here: [eldendle.vercel.app](http://eldendle.vercel.app/) https://preview.redd.it/2e60lsuhj4tf1.png?width=1899&format=png&auto=webp&s=fe847199a83d7bac0843eb9a74736be8cb89fd70 https://preview.redd.it/spvxj2cjj4tf1.png?width=1299&format=png&auto=webp&s=7d2ae480fa553355057a00d76e3ff2df7658f9f1
Pick your Vercel alternative only after weighing the pros and cons This sub has had many posts suggesting Vercel alternatives in the last few days. While some suggestions have been solid, others have been outright wrong. IMO it is super-vital to think through each alternative's benefits and limitations before choosing since hosting can get complicated to migrate. * Netlify - DDoS protection and WAF aren't included in non-enterprise plan. On a serverless offering, this can cause billing shocks. * Cloudflare - Nice for SSG and CDN pricing is awesome. But for SSR - Cloudflare Workers run on V8 runtime (and not Node) so every library that works on Node may not readily work. * Self-hosted VPS with Coolify (my preferred choice) - Best budget-wise, no platform locking, but needs initial build & deploy setup. * Railway - Nice predictable pricing, good build & deploy DX, doesn't offer CDN so need to combine with something like Cloudflare. * AWS / GCP services - Make good sense if you are already using these cloud providers, otherwise overwhelming number of offerings and options. Choose wisely, fellas!
Built a web app to help tracking currency spent and share your crafted items. Hello there, I build a web app to track the currency you spent on crafting an item, and at the end it will give you the option to share your item after pasting an image of the crafted item. which will give you the image in the description. \- You can add some currencies to be your favorite, so that u dont have to switch tabs every now and then. Its a very simple app that I can use to share my crafts with my friends, feel free to use it as well <3 Looking for your feedback to make it better. [https://poe-2-crafting-tracker.vercel.app/](https://poe-2-crafting-tracker.vercel.app/)
Vercel Controversy: Ethics, Backlash, and a Migration Guide to Netlify
[Resource] Attendance Calculator | Fully automatic https://preview.redd.it/38f8qocjr1tf1.png?width=1160&format=png&auto=webp&s=c98d369dfdbd4b0205f3aa3d61e5b8e10c645b34 https://preview.redd.it/bb1qmx0lr1tf1.png?width=1142&format=png&auto=webp&s=395b61c481a2a01ff5b9057c05451dbc5aecfd86 https://preview.redd.it/g0sr5aumr1tf1.png?width=1207&format=png&auto=webp&s=cfaad1ad929ee54e7f3482ccb5927a132ba169c2 https://preview.redd.it/biaivw2pr1tf1.png?width=1128&format=png&auto=webp&s=97626a40c7f70d3d0d659dd8699f777792417f1c Website: [Click](https://pesuattendance.vercel.app/) Features: * Connects to your PESU Academy account. * Shows your current attendance for each subject. * Live-calculates how many classes you can miss with a slider. * Works for both RR / EC Campus * Automatically loads calender and holidays * User/Pass are not saved on the server ensuring no logging * Smooth Fluid UI **Appreciate feedback / suggestions** Will clean up the code and put it on GitHub soon. Hope it's useful.
A Civil Rights and ICE tracking tool. Resharing in light of all the increasing fascism The insanity of everything going on with this whole ICE gestapo bullshit is mind boggling. And with growing concerns for our civil liberties, having access to safe, reliable tools is more important than ever. I built Firewatch to empower and protect our communities, and I want to share it again for anyone who feels at risk. Especially since they are taking down other apps from mobile app stores. It is a free and open-source app providing real-time ICE raid alerts, a nationwide map of community-reported incidents, secure incident reporting, Know Your Rights education in multiple languages, legal-grade evidence recording, emergency response tools, and sanctuary resources—all fully private, locally stored, and accessible offline. Just add your own free Gemini API key. If you don't want to use the AI features (the main AI feature is the ability to search your local area for incident and community reports), no worries! 90% of the app will still function perfectly without it. No catch, completely legal, and protected by the 1st Amendment (well...at least it used to be 🙃). Available now to protect civil rights and keep communities safe: https://firewatch-ice.vercel.app This app is built on a foundation of total privacy: It is serverless. There is no central server that I or anyone else can access. It does not track you or store your data. All information is stored locally on your device and is under your control. It is 100% open-source. Every line of code is public on GitHub for anyone to inspect and verify: https://github.com/severian42/Firewatch Key Features: - Real-Time ICE Raid Alerts: Stay informed with community-reported alerts. - Know Your Rights: Essential legal information available in multiple languages, accessible offline. - Secure Incident Reporting & Evidence Recording: Document events safely with legal-grade evidence tools. - Emergency Response & Sanctuary Resources: Quickly access help and find safe locations. Stay safe, know your rights, and support your community. App: https://firewatch-ice.vercel.app Code: https://github.com/severian42/Firewatch Side note: Also, if anyone has any other ideas for apps that could be built as tools to fight back and help protect each other, please drop them in the comments. I want to build more tools for the communities and people who need it. It seems that we, as the civilian population, are the only ones willing to do anything about this administrations insanity on the ground level
Best 8 Free AI Image Generators shared this week Thanks to everyone that posted a free AI image generator this week. I have tested all the links that were shared this week about Free AI Image generators. It is amazing that so many of them exist and with high quality and fast generations. Therefore I put together the list of best ai image generators found on this subreddit this week below. It will save you a lot of money and the time you spend searching for Free AI image generator is greatly reduced. Please let me know if I missed anyone mentioned this week. or comment below **The list can be found below:** 1. [https://aifreeforever.com/image-generators](https://aifreeforever.com/image-generators) \- Top quality free AI images with good prompting. offers unlimited AI Image generation per day without signing up nor login in. Contains some ads. Contains 120 seconds time before next generation. Quality is amazing. slightly can process adult nsfw prompts but converts it to bikinni clothing. Text generation is average. 2. [https://freegen.app/](https://freegen.app/) \- Completely free with no signup required. Instant access to AI art creation. Simple and straightforward interface 3. [https://eyedance.net/es](https://eyedance.net/es) \- Truly unlimited free generations with zero login requirements. Spanish interface but works internationally. Uses Flux Krea and Flux Kontext models. Zero data retention - complete privacy. Takes 10-30 seconds per image. Supports multiple formats and sizes up to 1024x1024. Has image editing features too. Works great on mobile. You own all generated images. 4. [https://pollinations.ai/](https://pollinations.ai/) \- Free open-source API platform. Requires JavaScript enabled. No local hardware needed - everything's cloud-based. 5. [https://visualgpt.io/ai-models/nano-banana](https://visualgpt.io/ai-models/nano-banana) \- Most likely free tier and definitely not unlimited but quality is high. you can generate without signup. i dont know their limit. you can comment when you test. Really fast and consistent results. Great for backgrounds, details, and portrait enhancements. Saves hours compared to Photoshop. Easy for non-technical users. Works within VisualGPT's larger tool ecosystem. 6. [https://dreamator-ai.vercel.app/](https://dreamator-ai.vercel.app/) \- Hosted on Vercel platform - means it's fast and modern. Clean minimalist interface. Likely offers free tier or demo access. Global CDN makes it quick worldwide. Focuses on simple text-to-image generation. Reliable uptime and automatic scaling. 7. [https://www.createimg.com/](https://www.createimg.com/) \- In addition to image generation, it has AI Background Changer for swapping backgrounds. AI Cloth Changer lets you try different outfits. AI Hairstyle to test different hair looks. Colorize Photo for old black and white images. AI Girl for character generation. AI Cartoonizer turns photos into cartoon style. More editing-focused than creation-focused. 8. [https://stable-diffusion-web.com/models](https://stable-diffusion-web.com/models) \- SD 3 Medium for faster experiments. SD 3.5 Large for high-end professional work. Great for learning which model fits your needs. Helps understand speed vs quality tradeoffs. Let me know if you want me to review test and add your image generator to the list. If you have tested and have coments on any one of them please comment below.
Developers drop Vercel, call for boycott after CEO posts selfie with Netanyahu
Codex so far I just upgraded to the Pro plan recently. This is unrelated but ChatGPT Pro with MCPs really feel like it’s able to give novel ideas and find breakthrough research. Anyway, I’ve been coding day and night with Codex and limits are nowhere to be seen which is great. But holy sh*t MCPs with Codex are just absolutely NUTS. I’ve been using different MCPs from Smithery and it’s been really useful so far. But this is where it gets me - I was thinking - hm my Vercel build isn’t working for this new project it’s kinda frustrating. It’s not a mini project I wanna give this much time to for something so simple of an issue. I spent a few mins and I was like you know what? Why if I let codex figure this out. Minute 1 -> Find Vercel MCP Minute 2 -> Add it to Codex very easily Minute ~5 -> Codex is like here’s what I found after scouring through the entire project in Vercel (build logs, deployments, etc), go change this setting And it worked!! Absolutely flawless. What I’m trying to say is, the ‘method’ of doing things is so much more efficient now. *As long as you have the security in mind* (like I excluded the deploy to Vercel function in the mcp), you can get stuff done 500x better than your competitor who doesn’t want to/know how to leverage these. Thank you for everyone who made Codex possible.
Built a little browser game using NOTHING BUT AI tools You can check it out here: [https://the-best-game-in-the-world.vercel.app/](https://the-best-game-in-the-world.vercel.app/) Took just a couple hours of very interrupted work. I used Claude Opus 4.1 and it basically did all of the code. The 3d assets (car and house) were text to 3d model generated using 3daistudio... you could probably prompt out better models than me honestly. Done using threejs, you can give it a try on mobile or desktop, both work. What do you guys think? Is anyone here working on anything remotely similar?
ELDENDLE FIRST UPDATE! Hey guys! 👋 Thank you all for the feedback on the first post. I’ve read every comment and suggestion, and I’m already working on bigger updates based on what you shared. This first patch is a small one, focused on community requests and quality-of-life improvements to make the experience smoother. **✨ What’s new:** * Added **orange feedback for partial matches!** * Improved the win modal, **with a button to switch directly between challenges.** * Added attempt counter for unlocking hints. * Added a timer for the next challenge at the top of the splash page. * Fixed duplicate input bug (pressing Enter with empty input). * Added new Boss Challenge category: **Story Relevance.** (Story Relevance indicates how important a boss is to the main storyline. **“Mandatory”** bosses are required to progress, while **“Optional”** ones are side encounters.) * Improved data accuracy for multiple bosses. * General polish and stability improvements. * Added **7 new characters**: * Commander Niall * Ancestor Spirit * Tibia Mariner * White Mask Varré * Melina * Miriel, Pastor of Vows * Smithing Master Hewg More updates are on the way soon, including new modes and features inspired by your feedback! Thanks again for playing and supporting Eldendle 💛 Play here: [eldendle.vercel.app](http://eldendle.vercel.app/) https://preview.redd.it/2e60lsuhj4tf1.png?width=1899&format=png&auto=webp&s=fe847199a83d7bac0843eb9a74736be8cb89fd70 https://preview.redd.it/spvxj2cjj4tf1.png?width=1299&format=png&auto=webp&s=7d2ae480fa553355057a00d76e3ff2df7658f9f1
This VibeCheck is powered by Pluggo — your AI-powered social listening platform. We find high-intent leads from Reddit, Twitter, Hacker News, and more, so you can engage prospects when it matters most.
Check the vibe of any brand and see how people really feel about them online.
Check Another BrandLet others see how Vercel is perceived online
Everything you need to know about finding your next customer on social media.
Still have questions? We'd love to help!
Reach out at hello@pluggo.ai
Get real-time alerts when sentiment shifts, competitors are mentioned, or new opportunities arise. Track your brand's vibe over time with Pluggo's continuous social listening.