Forem: Jawuil Pineda

AstroKit: Astro's Boilerplate for Launching Your Next Project in Record Time 🚀

Jawuil Pineda — Mon, 23 Feb 2026 16:00:47 +0000

If you're a web developer or an indie hacker, you probably know the thrill of having a new idea for a micro-SaaS or a side project. But you also know the most tedious part: the first 3 or 4 hours (sometimes days) configuring the environment, the framework, basic integrations, and folder structure.

As developers, we want to write our product logic and ship quickly, not reinvent the wheel by configuring the same thing over and over again.

To solve this problem, today I present AstroKit.

🛑 The Problem: Initial Friction
Every time I started a new project, I found myself repeating the same process:

Install the framework.

Configure the linter and code formatter.

Integrate Tailwind CSS (or your favorite styling tool).

Prepare the basic component structure and SEO.

That initial friction is the number one enemy of productivity and creative drive. In the minimalist entrepreneurship philosophy, speed to launch is everything for validating an idea.

⚡️ The Solution: What is AstroKit?
AstroKit is an Astro boilerplate designed specifically for builders and creators who value their time. It packages best practices and modern tools into a single, ready-to-use template.

The goal is simple: to go from idea to code in minutes.

Why choose Astro?
Astro has become one of the best options for building fast websites, blogs, portfolios, and SaaS landing pages thanks to its Island architecture and zero-JavaScript by default. By combining Astro with a robust boilerplate, you get:

Extreme performance (Perfect scores in Lighthouse).

Flawless development experience (DX).

Flexibility to use your favorite React, Vue, or Svelte components if needed.

🚀 We're live on Product Hunt!

Launching a product is always a huge milestone, and today AstroKit is officially on Product Hunt.

If you're a maker, enjoy building in public, or are simply looking to optimize your workflow for your next project, I'd love for you to check it out. Your technical feedback and support mean a lot to me in continuing to improve this tool.

👉 Support AstroKit on Product Hunt here : https://www.producthunt.com/products/astrokit

💬 Let's talk! I want to know what other developers on Dev.to think: What's the one tool or configuration that should always be in your initial stack when starting a project from scratch?

Leave it in the comments, and I'll see you on Product Hunt! 🚢

markdown-for-agents-why-html-is-dying-for-ais

Jawuil Pineda — Sun, 15 Feb 2026 20:24:43 +0000

We're still obsessed with Core Web Vitals and traditional human SEO. But there's an uncomfortable reality growing in web traffic: AI Agents.

Unlike users, AI agents (like those from OpenAI, Anthropic, or Perplexity) don't care about your CSS or animations. They care about structured data. And unfortunately, we're feeding them that data in the worst possible format: heavy, messy HTML.

Cloudflare just launched Markdown for Agents, a solution that promises to be a game-changer for "AI SEO." Here's what it is, why you should enable it, and how it impacts your LLM costs.

The Problem: Your HTML Is "Garbage" for an LLM

Feeding a language model with raw HTML is inefficient. A <nav> tag, dozens of nested divs, and tracking scripts have no semantic value for a bot. It's noise.

According to data published by Cloudflare, an average blog post weighs 16,180 tokens in HTML. That same content, converted to Markdown, weighs only 3,150 tokens.

We're talking about an 80% reduction in token usage.

If you're building AI applications or want your content to be cited by them, this metric is worth its weight in gold. Fewer tokens mean faster responses and less computational cost.

What is Cloudflare Markdown for Agents?

It's a new feature that enables real-time content conversion at the edge.

Instead of forcing the agent developer to scrape and clean your HTML (using their own resources), the Cloudflare network detects the request and transforms your HTML into clean Markdown before sending it.

The result is an optimized document, free of the "dead weight" of web design, ready to be consumed by an LLM.

How to Implement Markdown for Agents (Tutorial)

The implementation is based on standard content negotiation. You don't have to install anything on your server; everything happens in the Cloudflare layer.

Step 1: The Request

The AI agent (or you, if you're testing) must include a specific header in the HTTP request:

Accept: text/markdown

Step 2: The Response

Cloudflare intercepts this header. If the functionality is enabled in your region, it will retrieve the HTML from your source, convert it, and return Markdown.

You can try it right now with this command in your terminal against the Cloudflare documentation:

curl [https://developers.cloudflare.com/fundamentals/reference/markdown-for-agents/](https://developers.cloudflare.com/fundamentals/reference/markdown-for-agents/) \
-H "Accept: text/markdown"

The response also includes an x-markdown-tokens header that tells you the estimated number of tokens in the document, ideal for managing context windows in your applications.

Requirements and Availability: Is it free?

Here's the important detail for indie hackers and developers. Although the tool is powerful, it's not available for all plans at the moment.

Status: Open Beta.
Supported Plans: Pro, Business, and Enterprise.
SaaS: Also included if you're an SSL for SaaS customer.
Free Plan: Currently not included in the free plan.

To activate it, go to your Cloudflare dashboard, select your zone, and find the toggle under "Quick Actions."

Conclusion: The SEO of the Future is Semantic

If you have a SaaS project, technical documentation, or a blog and already pay for a Cloudflare plan, activate it today.

Markdown for AI tools (LLM Optimization) is the next big battle for traffic. If your content is easier for an agent to read than your competitors', you have a huge competitive advantage.

It's not just about code; it's about distribution. In 2026, making life easier for bots is a smart business strategy.

Source: https://blog.cloudflare.com/markdown-for-agents/

I work in agriculture by day and code AI by night. Roast my first SaaS MVP!

Jawuil Pineda — Sun, 07 Dec 2025 01:23:15 +0000

Hi Indie Hackers! 👋

My name is Jawuil, I'm a 21-year-old developer from Venezuela.

The Context: My daily life is quite different from the typical Silicon Valley dev. I live in a rural area and work in agriculture on my family's farm during the day. But every night (and whenever I have a break), I'm glued to my laptop learning to code and building products. I started my coding journey seriously around mid-2024.

The Problem: I noticed that generic prompts don't work everywhere. Trying to use a simple ChatGPT prompt inside complex tools like Cursor (for code), v0.dev, or Jasper usually gives average results. Each AI needs a specific format or structure to work correctly.

The Solution: I built https://accelprompt.xyz/

It is a specialized prompt generator designed to create optimized instructions for specific AI tools. It supports:

Coding: Cursor, v0.dev, Lovable.

Writing: Jasper, Copy.ai.

Visuals: DALL-E, Sora (Video).

It handles the formatting so you just focus on the idea, and it generates the precise instruction the AI needs.

The "Boring" Stack: I believe in shipping fast, so I kept it simple:

Backend: Vanilla PHP (Yes, really!) & Supabase for Auth.

Frontend: Tailwind CSS & Vanilla JS.

Payments: I'm using Ko-fi right now to keep fees low while I validate the idea. It's not fancy, but it works.

Where I need your help: To be honest, I haven't done any real marketing yet. This post is my first time sharing it publicly. Marketing is my biggest weakness right now, and I often feel scattered trying to promote my work while balancing my day job.

Roast my landing page: I just updated the H1 to be more specific. Is the value proposition clear now?

Pricing/Stack: Do you think using Ko-fi for a SaaS is a dealbreaker for users, or is it fine for an MVP?

Thank you so much for reading. The Indie Hackers community has been a huge inspiration for me to keep pushing despite the challenges here.

Cheers! Jawuil

I work in agriculture by day and code AI by night. Roast my first SaaS MVP!

Jawuil Pineda — Sun, 07 Dec 2025 01:15:23 +0000

Hi Indie Hackers! 👋

My name is Jawuil, I'm a 21-year-old developer from Venezuela.

The Solution: I built https://accelprompt.xyz

It is a specialized prompt generator designed to create optimized instructions for specific AI tools. It supports:

Coding: Cursor, v0.dev, Lovable.

Writing: Jasper, Copy.ai.

Visuals: DALL-E, Sora (Video).

It handles the formatting so you just focus on the idea, and it generates the precise instruction the AI needs.

The "Boring" Stack: I believe in shipping fast, so I kept it simple:

Backend: Vanilla PHP (Yes, really!) & Supabase for Auth.

Frontend: Tailwind CSS & Vanilla JS.

Payments: I'm using Ko-fi right now to keep fees low while I validate the idea. It's not fancy, but it works.

Roast my landing page: I just updated the H1 to be more specific. Is the value proposition clear now?

Pricing/Stack: Do you think using Ko-fi for a SaaS is a dealbreaker for users, or is it fine for an MVP?

Thank you so much for reading. The Indie Hackers community has been a huge inspiration for me to keep pushing despite the challenges here.

Cheers! Jawuil

Free tool to remove watermarks from AI images, 100% private

Jawuil Pineda — Sat, 29 Nov 2025 21:00:07 +0000

I've created a tool to remove watermarks from images.
Perfect for those who experiment with AI and create many images with it.
It's FREE and 100% private; your images are not sent to any server. Everything is processed in your browser thanks to OpenCV technology.

https://jawuil.dev/remove-watermark

Tool to create free n8n flows without login (in beta)

Jawuil Pineda — Tue, 04 Nov 2025 18:28:13 +0000

Hello everyone!

I'm creating a tool to help you build n8n flows using AI, and I'd appreciate any feedback or suggestions.

Here's the link; it's free to use, no registration required.

https://jawuil.dev/n8n-flow