Forem: Yash Garudkar

I built a 1.73 KB axios alternative with zero dependencies - glyde on npm

Yash Garudkar — Sun, 17 May 2026 09:54:34 +0000

I got mass an email from npm in March 2026 about the axios supply chain attack. A North Korean state actor compromised a maintainer and injected malicious code into the most popular HTTP client on npm. Millions of projects affected overnight.

That's when I looked at what axios actually does: it merges config, builds URLs, adds headers, runs interceptors, and wraps the response. All of that is maybe 200 lines of TypeScript on top of native fetch. But axios ships 53 KB gzipped with 2 dependencies that have their own dependencies.

So I built glyde - a TypeScript-first HTTP client with zero runtime dependencies. 1.73 KB gzipped.

What it looks like

import plane from "glyde"

const api = plane({ baseURL: "https://api.example.com" })

// Typed GET
const { data } = await api.get<User[]>("/users")

// POST with body
await api.post("/users", { name: "Yash", role: "admin" })

// Query params
await api.get("/search", { params: { q: "glyde", page: 1 } })

plane() creates an independent instance with its own config and interceptors. No shared global state.

Typed errors instead of status code checking

This is what error handling looks like with most HTTP clients:

// The old way
try {
  await axios.get("/data")
} catch (err) {
  if (err.response?.status === 404) { /* maybe? */ }
  // What type is err? Who knows.
}

With glyde, errors have a type hierarchy with type guards:

import { isHttpError, isTimeoutError, isGlydeError } from "glyde"

try {
  await api.get("/data")
} catch (err) {
  if (isHttpError(err)) {
    // TypeScript knows: err.status, err.response, err.config
    console.log(err.status)        // 404
    console.log(err.response?.data) // parsed body
  }

  if (isTimeoutError(err)) {
    // request exceeded timeout
  }
}

The error hierarchy:

GlydeError (base)
+-- HttpError      - non-2xx response (has status, response, config)
+-- TimeoutError   - request exceeded timeout
+-- NetworkError   - fetch failed (DNS, offline, CORS)

Async interceptors

Unlike most libraries that only support synchronous transforms, glyde interceptors are fully async:

// Refresh a token before every request
api.interceptors.request.use(async (config) => {
  const token = await getToken()
  return {
    ...config,
    headers: { ...config.headers, Authorization: `Bearer ${token}` },
  }
})

// Unwrap nested API responses
api.interceptors.response.use((response) => ({
  ...response,
  data: response.data.result,
}))

Next.js App Router pattern

glyde works anywhere fetch exists, but I designed it with Next.js in mind. The recommended pattern:

Server-side (tower):

import plane from "glyde"
import { cookies } from "next/headers"

export async function tower() {
  const api = plane({ baseURL: process.env.API_BASE_URL })
  const cookieStore = await cookies()

  api.interceptors.request.use((config) => {
    const token = cookieStore.get("access_token")?.value
    if (token) {
      config.headers = { ...config.headers, Authorization: `Bearer ${token}` }
    }
    return config
  })

  return api
}

Client-side (passenger):

"use client"
import plane from "glyde"

export const passenger = plane({ baseURL: "/api/proxy" })

passenger.interceptors.response.use(
  (response) => response,
  (error) => {
    if (error?.status === 401) window.location.href = "/login"
    throw error
  }
)

Token refresh? Handle it in Next.js middleware where you can actually write cookies. Don't fight the framework.

The numbers

	glyde	axios
Size (gzipped)	1.73 KB	53 KB
Dependencies	0	2
TypeScript	Written in TS	Types bolted on
Engine	Native fetch	Legacy XHR

Install

npm install glyde

Works in browsers, Node.js 18+, Bun, Deno, and Cloudflare Workers.

I'm actively working on retry plugins, request deduplication, and more interceptor examples. If you have feedback or feature requests, open an issue on GitHub.

Behind the Scenes: Volunteering at AWS Community Day Mumbai 2025

Yash Garudkar — Wed, 15 Oct 2025 17:22:59 +0000

There’s something special about community events — the late nights, last-minute fixes, endless coffee, and the sense of purpose that keeps everyone moving.
For me, AWS Community Day Mumbai 2025 was all of that — and more.

The Journey Begins

It started with a drive to Mumbai — a backpack full of essentials, planning to-do lists, and a calendar day packed tighter than a Lambda cold start. The days leading up to the event were a mix of excitement and execution.

We were up late into the night preparing ID cards for everyone attending the event venue, and ensuring that every detail — from speaker kits to attendee goodies — was ready for the big day. The venue, the volunteers, the coordination & communication — everything had to click perfectly.

Prepping for the D-Day

From checking registrations and packing swags, to supporting speakers with their setup and helping sponsors with their booths — it was all hands on deck.

It’s easy to underestimate the amount of effort that goes into a large-scale tech event, but when you’re there, watching ideas turn into action, it’s pure magic.

A fun story: I was at the check-in desk in the first half. There was one surprise we came across. The ID cards weren't alphabetically sorted. I went Eagle mode with my eyes and laid ID cards piles on my desk. Now that I had seen names, it was easier to estimate which card was in what pile. Sounds hard, was fun😂.

The Event Unfolds

The day began early — 7:00 AM — with volunteers setting up registration counters and preparing the hall for the crowd. By 9:00 AM, the venue was buzzing with builders, engineers, and cloud enthusiasts ready to dive into a day packed with learning and connection.

The morning keynotes set the tone perfectly.

Luis Guirigay (AWS) spoke about architecting for reliability and efficiency at scale — a topic that resonated with every backend engineer and solutions architect in the room.

James Conway (Temporal) followed up with a session on building AI-ready workflows using Temporal — bridging reliability and scalability in a way that felt refreshingly practical.

Swaminathan R (GitLab) rounded off the keynote block by exploring AI-native software development — an insight into how developer tools themselves are evolving.

A quick quiz competition kept the crowd energised before we broke for a much-needed tea and networking session. Conversations flowed naturally — from architecture discussions to startup ideas, and of course, a few “which AWS service do you love/hate most” debates.

Post-break, the event split into multiple tracks, covering everything from Elastic’s conversational agent builder, IoT data management, and real-time AWS pipelines, to agentic AI workflows by Ekta Shah (Morgan Stanley) and Shubham Agnihotri (IDFC First Bank).

Meanwhile, a hands-on workshop on AI-Driven Agents for Enterprise Workflows by Deepak M gave developers a chance to experiment live — one of the most engaging parts of the day.

The afternoon sessions showcased the technical depth of the community. Talks by Kunal Das (Cast AI) on zero-downtime migrations, Agilandeswari Sivasubramanian (NASDAQ) on cloud transformation, and Sreedharma Vijayan (TiDB) on distributed databases showed just how much innovation is happening across industries.

By lunchtime, the hall was buzzing again — this time around the open networking lunch setup. The spread was delicious, and the conversations even better. Developers, speakers, and sponsors mingled freely, discussing AI and dessert.

The post-lunch sessions featured some standout moments — Jones Zachariah Noel N (AWS Hero) deep-diving into AWS Lambda PowerTools, Bhuvaneswari Subramani (AWS Hero) discussing AI-driven developer productivity, and Kunal Das returning with an advanced workshop on Kubernetes automation that pushed everyone to think about scaling smartly.

Talks from Shivay Lamba (Couchbase), Vivek Raja (AWS Hero), and Rohan Khameshra (OLake by Datazip) kept the AI and data-focused momentum strong.

The late-afternoon highlight for me was Donnie Prakoso (AWS) — his talk on the future of microservices with agentic communication wasn’t just technical; it was visionary. Soon after, Manoj Fernando (AWS Hero) and Nilesh Vaghela (AWS Hero) brought in their deep technical experience to the stage — both humble, insightful, and incredibly engaging to speak with later.

By 4:30 PM, attendees enjoyed a short break before the final sessions rolled out. The day wrapped up beautifully with Ridhima Kapoor (AWS) delivering her keynote on The Power of Community — the perfect message to end a day built by, and for, the community itself.

And then came the best part — the high tea networking session.
The atmosphere was relaxed, filled with laughter, handshakes, and LinkedIn exchanges. Conversations about the day’s topics transitioned into future ideas, collaborations, and community goals.

Interacting with Jones Zachariah Noel (AWS Hero), Donnie Prakoso (AWS), Nilesh Vaghela (AWS Hero), and Manoj Fernando (AWS Hero) was truly insightful. Hearing their experiences on stage — and later, chatting with them off-stage — was both humbling and motivating.

It wasn’t just an event; it felt like a reunion of passionate builders who understood what it means to create, share, and grow together.

Teamwork That Makes It Happen

Community events like this don’t happen by accident. They’re the result of months of preparation and a passionate team.

The AWS User Group Mumbai organising team, led by Sanchit Jain, along with Abhishek Wagh, Atharva Rode, Angel Arora, Nilesh Shinde, Abhishek Maurya, Jayesh More, Akash Chavan, Ritik Chaturvedi, Chetan Bhasney, Darshan Wagh, Pushkar Thakur, Samruddhi Borse, and Yash Pimple, worked tirelessly to make it all seamless.

Why Community Matters

Volunteering isn’t just about managing logistics — it’s about connection. You meet people who share your passion for tech, who give their weekends to help others learn, and who make the ecosystem stronger through contribution.

Being part of AWS User Group Mumbai reminded me once again that community isn’t built by companies — it’s built by people.

And that’s what keeps us all coming back.

Here’s to more learning, more collaboration, and more AWS magic in the years to come.

LinkedIn: https://linkedin.com/in/yashgkar