Forem: Jithun Methusahan

I’m finally doing it. AI-KIT goes live in 5 days. 🚀

Jithun Methusahan — Fri, 23 Jan 2026 11:56:41 +0000

I’ve spent the last 1 month solving a problem that’s been driving me crazy "chatbots".

I realized that if I’m struggling with this, others probably are too. So, I decided to build a solution and, more importantly, open-source it.
In exactly 7 days, I’ll be flipping the switch on the GitHub repo.

What is AI-KIT?
It’s a premade ui and api intergration ready tool designed to help designers, businesses, and developers to integrate a chatbot to their website without coding knowledge or repetitive work.

The Goal:
I’m not just launching a tool. I’m looking for a community of builders to help shape its future.

Want to be the first to contribute:
Drop a "🚀" in the comments.
Follow me here for daily updates on the final countdown.
GitHub: https://lnkd.in/ghnG7TcM

hashtag#OpenSource hashtag#BuildingInPublic hashtag#SoftwareEngineering hashtag#Launch

A Beginner's Guide to APIs: Supercharge Your Apps with 100+ Free APIs

Jithun Methusahan — Fri, 23 Jan 2026 11:55:42 +0000

A Beginner's Guide to APIs: Supercharge Your Apps with 100+ Free Integrations

APIs, or Application Programming Interfaces, let software applications communicate and share data or functionality. They act as intermediaries, so developers can access features from services like weather apps or payment processors without building everything from scratch.

Key Concepts
APIs use HTTP methods like GET to retrieve data, POST to create, PUT to update, and DELETE to remove resources. Most modern APIs follow RESTful principles and return data in JSON format for easy parsing in languages like Python or JavaScript. Authentication often uses API keys or OAuth tokens to secure access.

Usage Tips:

1.Read documentation first: Always review the official docs for endpoints, parameters, and examples to avoid trial-and-error.

2.Handle errors gracefully: Check status codes (e.g., 200 for success, 404 for not found) and implement try-catch blocks or retries.

3.Respect rate limits: Avoid exceeding request quotas by adding delays or caching responses.

4.Secure keys properly: Store API keys in environment variables, never in code, to prevent exposure.

5.Test with tools: Use Postman or curl to experiment before integrating into your projects.

6.Monitor usage: Track calls to prevent surprises like billing overages.
Your Repo Resource

Repo Resource

I have a repository with hundreds of APIs as a hands-on resource for experimentation. Clone it locally, explore endpoints with Python scripts using the requests library or JavaScript with fetch, and build small projects like data aggregators to apply these tips. because the more you practise, the more you learn

https://github.com/public-apis/public-apis

A Beginner's Guide to APIs: Supercharge Your Apps with 100+ Free APIs

Jithun Methusahan — Sun, 04 Jan 2026 15:40:27 +0000

A Beginner's Guide to APIs: Supercharge Your Apps with 100+ Free Integrations

Usage Tips:

1.Read documentation first: Always review the official docs for endpoints, parameters, and examples to avoid trial-and-error.

2.Handle errors gracefully: Check status codes (e.g., 200 for success, 404 for not found) and implement try-catch blocks or retries.

3.Respect rate limits: Avoid exceeding request quotas by adding delays or caching responses.

4.Secure keys properly: Store API keys in environment variables, never in code, to prevent exposure.

5.Test with tools: Use Postman or curl to experiment before integrating into your projects.

6.Monitor usage: Track calls to prevent surprises like billing overages.
Your Repo Resource

Repo Resource

https://github.com/public-apis/public-apis

Stanford's AI Lab analyzed hundreds of coffee meetings here are 5 astonishing takeaways that every AI builder needs to know!

Jithun Methusahan — Fri, 02 Jan 2026 14:54:44 +0000

When I first skimmed the paper from Stanford's AI Lab, I thought it was some groundbreaking novel. Nope—just raw insights from hundreds of casual coffee meetings with top AI researchers.

As a young AI builder, it hit hard. We're out here making hilarious AI-generated videos (shoutout to the memes), while the world races ahead with real breakthroughs. Most folks still imagine AI as chatbots or personal assistants... but this paper reveals what's actually brewing.

I broke it all down in this detailed summary—perfect for builders like us chasing the next big thing

Read the full breakdown here👇
https://jithunmethusahan.medium.com/stanfords-ai-lab-held-hundreds-of-meetings-over-coffee-here-are-5-astonishing-takeaways-6d84041a1a7e

Quick 5 Takeaways to Fuel Your Next Project:

Hidden Collaboration Goldmines: Informal chats spark 80% of breakthrough ideas—skip the formal meetings, grab coffee.
The "AI Winters" Myth: Cycles of hype and bust? Not anymore—steady progress is the new normal.
Ethics Isn't a Checkbox: Real-world AI dilemmas get solved over lattes, not in boardrooms.
Interdisciplinary Magic: Blend AI with bio/neuro for moonshots—solo coding won't cut it.
Global Race Reality: Top labs are 10x ahead; us builders need open-source hustle.
This stuff changed how I approach my ML projects. What's your biggest AI "aha" from casual convos?

How I built a serverless image converter using React & WebAssembly

Jithun Methusahan — Thu, 25 Dec 2025 05:55:06 +0000

Hey everyone!
I wanted to learn how to manipulate images in the browser without a backend. So I built Pixel Perfect Convert.
It uses the HTML5 Canvas API to convert WebP/HEIC to PNG. No files are uploaded to any server.
Stack: React, Tailwind, JSZip.
You can try the demo here: https://pixel-perfect-convert.vercel.app/
I launched on Product Hunt today if you want to support: https://www.producthunt.com/products/pixel-perfect-convert

The Hidden Dockerfile Mistakes That Waste Hours (Thinking of Building a Tool for This)

Jithun Methusahan — Tue, 02 Dec 2025 03:09:58 +0000

Developers love to pretend their Dockerfiles are perfect until a build takes eight minutes and the image balloons bigger than a AAA game update. Most of the time, the cause is embarrassingly simple stuff: missing version pins, unnecessary layers, or copying half the universe into the image.

Example mistakes I keep seeing:

"RUN apt-get install python

No version pin. Enjoy the surprise upgrades later.

COPY . .

Congrats, you just shoved your entire repo and node_modules into the image because why not.

RUN pip install --upgrade pip
RUN pip install -r requirements.txt

Two separate layers doing what one should handle.

I’m exploring a small, focused tool that analyzes Dockerfiles and built images to catch these issues quickly. No AI circus, just rule checks and clear suggestions.

I’m curious how common this pain actually is.
Would something like this be useful in your workflow, or do you already rely on another tool?

Feel free to comment your thoughts or examples of Dockerfile mistakes you’ve faced