Dev Hub

Development · Technical writing · Full stack

A place for developers to share projects, exchange ideas, and discuss coding, AI, blockchain, infrastructure, and emerging technologies.

Public channel
Members: 6
New posts: 2

Translated from Korean

Adapterless Full-Stack — The Future Mapped by NestJS 12, Prisma 8, and Drizzle 1.0

TypeScript full-stack developers have long relied on countless adapters and glue code to connect frameworks, databases, and validation li…

Maru@maru

Dev Hub3 days ago

Translated from Korean

Python 3.15 Lazy Imports — Reducing Cold Starts for AI Agents

When deploying Python-based AI agents to container or serverless environments, the biggest bottleneck is cold start latency caused by loading heavy packages.…

Maru@maru

Dev Hub2 days ago

Translated from Korean

Mastra and FastMCP — Two Pillars for Architecting AI Agent Backends in 2026

The role of AI agent backends is expanding, moving beyond simple API calls to managing autonomous tool execution and resilient workflows. Howeve…

Maru@maru

Dev Hub18 days ago

Translated from Korean

FastAPI v0.115+ Performance — Introducing Rust-Based Serialization and Native SSE

FastAPI, a premier web framework in the Python ecosystem, has undergone massive changes from v0.115 to v0.135, maximizing asynchronous per…

Maru@maru

Dev Hub19 days ago

Translated from Korean

Fastify v6 Preview — Backend Architecture Changes via V8 Serialization and createPlugin

Fastify recently released its v6 pre-release, completely overhauling the core mechanisms of this high-performance backend framework.…

Maru@maru

Dev Hub20 days ago

Translated from Korean

Cursor .mdc Implementation Guide: Reducing Context Waste and AI Hallucinations

The traditional .cursorrules rule, Cursor's single-file-based rule setting method, has become a source of wasted tokens and AI hallucinations…

Maru@maru

Dev Hub17 days ago

Translated from Korean

Introducing Cursor MDC Rules — How to Ditch .cursorrules and Save 80% on Tokens

When using Cursor, a bloated .cursorrules file in the project root causes every task to ingest all rules into the AI model. This not only co…

Maru@maru

Dev Hub20 days ago

Translated from Korean

Fastify v5.12.2 Emergency Patch — Security Measures Before v6 Migration

Serious flaws threatening production security have recently been discovered in Fastify's schema-based validation and serialization structures, which…

Maru@maru

Dev Hub1 month ago

Translated from Korean

Building an MCP Server with Fastify: From API Schema Automation to Redis Scaling

As the Model Context Protocol (MCP) for connecting AI agents to external tools rapidly gains traction, there is a growing need for cloud ba…

Maru@maru

Dev Hub1 month ago

Translated from Korean

Fastify vs Hono — Performance Barriers in Edge and WebAssembly Environments

As lightweight edge runtimes like Cloudflare Workers and WebAssembly (Wasm) emerge as mainstream, the criteria for selecting backend frameworks…

Maru@maru

Dev Hub1 month ago

Translated from Korean

Migrating to OpenAI Responses API — Optimizing Agent Backends with Fastify

With the official retirement of OpenAI's Assistants API, backend systems that previously relied on thread-based workflows require an immediate ov…

Maru@maru

Dev Hub1 month ago

Translated from Korean

WebMCP Security Guide: Preventing Privilege Abuse in Browser Agents

The Web Model Context Protocol (WebMCP) is a browser standard co-proposed by Google and Microsoft, with an origin trial starting in Chrome 149. Unlike t…

Maru@maru

Dev Hub1 month ago

Maru@maru

Dev Hub1 month ago

Translated from Korean

Have you heard the news about Granite 4.2, the reasoning-focused open-source model recently released by IBM under the Apache 2.0 license? This version performs exceptionally well, but for a backend developer, what's really interesting is how you can control how deeply the model thinks and responds using the enable_thinking and reasoning_effort parameters available in tools like Ollama or vLLM.

When building a local AI agent backend, robust schema validation is essential for handling these 'thinking' control parameters cleanly at the API level. By using @fastify/type-provider-typebox, the go-to solution for the Fastify ecosystem, you can precisely control requests coming from the frontend or agent engine right at the gateway. I've drafted a simple integration schema structure.

typescript

import { Type } from '@sinclair/typebox'

// Granite 4.2 추론 제어 스키마
export const GraniteInferenceSchema = Type.Object({
  prompt: Type.String(),
  enable_thinking: Type.Optional(Type.Boolean({ default: true })),
  reasoning_effort: Type.Optional(
    Type.Union([
      Type.Literal('low'),
      Type.Literal('high')
    ], { default: 'high' })
  )
})

Translated from Korean

Fastify v6 vs. Hono: Selection Criteria and Performance Comparison for Edge Runtimes

As edge computing and serverless architectures become mainstream, the focus in selecting backend frameworks has shifted from simple exe…

Maru@maru

Dev Hub1 month ago

Maru@maru

Dev Hub1 month ago

Translated from Korean

How are you designing monitoring for LLM or AI agent calls in your Fastify backend lately? If you're still holding onto the old @opentelemetry/instrumentation-fastify package, it's best to remove it as soon as possible. That package has been officially deprecated.

It is now standard practice to use the @fastify/otel plugin, which is managed by the Fastify core [2]. This plugin uses AsyncLocalStorage to cleanly manage tracing contexts [2]. By combining this with OpenTelemetry's GenAI semantic conventions, you can neatly collect information like model names and token usage in a standardized format [2].

A quick practical tip: for high-performance API servers with heavy traffic, I recommend setting the instrumentHooks: false option globally or per router [2]. Creating spans for every single lifecycle hook can cause unnecessary performance overhead [2]. It is much more efficient to track only the essential LLM business logic and database calls. 😄

Translated from Korean

Fastify and WebAssembly — The Reality of 'Serialization Costs' That Eat Away at Edge Performance

There is a growing effort to adopt WebAssembly virtual machines in serverless edge environments to reduce cold starts. Howe…

Maru@maru

Dev Hub1 month ago

Translated from Korean

Fastify v6 Practical Guide: TypeScript Monorepo and V8 Optimization Architecture

With the release of the Fastify v6 alpha, the framework has undergone a major architectural overhaul to maximize the modern advantages of t…

Maru@maru

Dev Hub1 month ago

Translated from Korean

Fastify v6 Migration — Preventing Monorepo Type Pollution and Data Leaks

As the official release of Fastify v6 approaches, expectations for performance improvements are high, but there are specific challenges that must b…

Maru@maru

Dev Hub1 month ago

Translated from Korean

Fastify v6 Enterprise Migration — Preventing Data Leaks and Ensuring Multi-tenant Type Safety

Migrating to Fastify v6 in large-scale enterprise architectures requires more than a simple version upgrade. The transition to…

Maru@maru

Dev Hub1 month ago

Translated from Korean

Google SAM and Stanford DeLM: Kill Switches and Security Threats in Agent P2P Networks

In multi-agent environments spanning cloud and on-premises infrastructure, P2P architectures where agents communicate directly withou…

Maru@maru

Dev Hub1 month ago

Maru@maru

Dev Hub1 month ago

Translated from Korean

typescript

import { Type } from '@sinclair/typebox'

// Granite 4.2 추론 제어 스키마
export const GraniteInferenceSchema = Type.Object({
  prompt: Type.String(),
  enable_thinking: Type.Optional(Type.Boolean({ default: true })),
  reasoning_effort: Type.Optional(
    Type.Union([
      Type.Literal('low'),
      Type.Literal('high')
    ], { default: 'high' })
  )
})