Name: JSNation US 2025
Start: 2025-11-17T13:00:00.000Z
End: 2025-11-20T22:59:00.000Z

Lera

  Neha  Seth

Brian Whippo

Simon Ouyang

Daria Shabala

Seungho Park

Will Klein

Alexander Lichter

Karen Li

Alex Astrum

Richard Carrigan

Dipanshu Gupta

Varun

Amanda Witecki

Kyle Langford

Houda Mouttalib

Pascal Moser

Trust Jamin

James O'Reilly

Mauro Bartolomeoli

Kritiketan Sharma

Joana Santos

Noam Zingboim

Henry Lie

Francesco Sciuti

Benjamin Schachter

Herrington Darkholme

Elina Denisova

David Mytton

Leon Malisov

Maya Shavin

Hunor Márton Borbély

Mangirdas Darguzis

El rol del ingeniero de software está experimentando un cambio fundamental. A medida que AI se integra profundamente en nuestros flujos de trabajo, está surgiendo un nuevo paradigma: el Ingeniero de Software Nativo de AI. No se trata de ser reemplazado; se trata de amplificación. Esta charla presenta un manual práctico para que los desarrolladores hagan la transición a AI siendo un socio versátil en todo el ciclo de vida del desarrollo de software. Exploraremos el cambio de mentalidad crucial necesario para ver AI como un multiplicador de productividad, no una amenaza. Aprenderás cómo orquestar efectivamente agentes y herramientas de AI, incluyendo el nuevo Chrome DevTools MCP, para delegar tareas rutinarias y liberarte para enfocarte en la resolución de problemas de alto nivel, arquitectura e innovación. Iremos más allá de la generación básica de código y discutiremos la aplicación de AI al diseño de sistemas, pruebas, depuración y operaciones, transformando tu rol de escritor de código a pensador de sistemas. Únete a mí para descubrir cómo "confiar, pero verificar", mantener la propiedad y calidad última, y convertirte en un ingeniero nativo de AI indispensable que está dando forma al futuro del software.

El Ingeniero de Software Nativo de AI

Software engineering is evolving with AI and VIBE coding reshaping work, emphasizing collaboration and embracing AI. The future roadmap includes transitioning from augmented to AI-first and eventually AI-native developer experiences. AI integration in coding practices shapes a collaborative future, with tools evolving for startups and enterprises. AI tools aid in design, coding, and testing, offering varied assistance. Context relevance, spec-driven development, human review, and AI implementation challenges are key focus areas. AI boosts productivity but faces verification challenges, necessitating human oversight. The impact of AI on code reviews, talent development, and problem-solving evolution in coding practices is significant.

The role of the software engineer is undergoing a fundamental shift. As AI becomes deeply integrated into our workflows, a new paradigm is emerging: the AI-Native Software Engineer. This isn't about being replaced; it's about amplification. This talk presents a practical playbook for developers to transition to AI being a versatile partner across the entire software development lifecycle.We'll explore the crucial mindset shift required to see AI as a productivity multiplier, not a threat. You will learn how to effectively orchestrate AI agents and tools - including the new Chrome DevTools MCP - to delegate routine tasks and free yourself to focus on high-level problem-solving, architecture, and innovation. We will move beyond basic code generation and discuss applying AI to system design, testing, debugging, and operations, transforming your role from a code writer to a systems thinker. Join me to discover how to "trust, but verify," maintain ultimate ownership and quality, and become an indispensable AI-native engineer who is shaping the future of software.

The AI-Native Software Engineer

Addy Osmani

Arreglar errores de CORS es un rito de iniciación frustrante para cada desarrollador web. La API de Renderizado de Bun hace trivial ejecutar tu frontend y API en el mismo servidor, eliminando toda la categoría de errores de origen cruzado. Los desarrolladores ya no necesitan luchar con encabezados complejos del lado del servidor o solicitudes proxy solo para obtener datos de un dominio diferente.

Nunca te preocupes por CORS nuevamente

Talk on Bund: Overview of Bund, its speed optimization, components like runtime, package manager, test runner, and bundler. Efficient features include fast npm manager, pnpm lockfile support, Jest-compatible test runner, and optimized runtime components. Built-in features like modern password hashing, WebSocket server, file glob API, and database clients. API design focuses on common use cases and optimization. Enhanced JavaScript functionality with SIMD, native code, and performance improvements. Challenges in JavaScript productivity, BUN adoption, deployment, and support. React support, community involvement, security, Next.js compatibility, and transition.

Fixing CORS errors is a frustrating rite of passage for every web developer. Bun's Rendering API makes it trivial to run your frontend and API on the same server, eliminating the entire category of cross-origin errors. Developers no longer need to wrestle with complex server-side headers or proxy requests just to fetch data from a different domain.

Never Worry About CORS Again

Jarred Sumner

Model Context Protocol (MCP) introduce un enfoque estructurado para la gestión de contexto de LLM que aborda las limitaciones en los métodos de prompting tradicionales. En esta masterclass, aprenderás sobre el Model Context Protocol, su arquitectura y cómo construir y usar un MCP con Node.jsTabla de Contenidos:<ul><li>¿Qué es el Model Context Protocol?</li><li>Tipos de MCPs (Stdio, SSE, HTTP Streaming)</li><li>Comprendiendo Herramientas, Recursos y Prompts</li><li>Construyendo un MCP con el SDK Oficial de TypeScript en Node.js</li><li>Desplegando el MCP en la Nube (Heroku)</li><li>Integrando el MCP con Tu Herramienta de IA Favorita (Claude Desktop, Cursor, Windsurf, VS Code Copilot)</li><li>Consideraciones de Seguridad y Mejores Prácticas</li></ul>

Construir un MCP (Model Context Protocol) en Node.js

The Workshop covered various aspects of building Model Context Protocol (MCP) applications in Node.js with Heroku. Participants were instructed on setting up their environment, understanding MCP concepts, building MCP clients and servers, deploying chat applications, and integrating security measures. They also learned about configuring servers, managing resources, handling deployment challenges, and utilizing MCPs for data sources and security. The Workshop concluded by thanking the audience and sharing contact information.

Model Context Protocol (MCP) introduces a structured approach to LLM context management that addresses limitations in traditional prompting methods. In this workshop, you'll learn about the Model Context Protocol, its architecture, and how to build and use and MCP with Node.jsTable of Contents:<ul><li>What Is the Model Context Protocol?</li><li>Types of MCPs (Stdio, SSE, HTTP Streaming)</li><li>Understanding Tools, Resources, and Prompts</li><li>Building an MCP with the Official TypeScript SDK in Node.js</li><li>Deploying the MCP to the Cloud (Heroku)</li><li>Integrating the MCP with Your Favorite AI Tool (Claude Desktop, Cursor, Windsurf, VS Code Copilot)</li><li>Security Considerations and Best Practices</li></ul>

Build a MCP (Model Context Protocol) in Node.js

Julián Duque

Vamos a desmitificar los secretos de las herramientas modernas de JavaScript. En este masterclass, aprenderás cómo funcionan los linters, compiladores y transformaciones de código. Darás tus primeros pasos construyendo tus propias herramientas de desarrollo, escribiendo una regla de ESLint, un plugin de compilador y una transformación de código. Aprende estas habilidades para ayudar a tu equipo y profundizar tu comprensión de los lenguajes de computadora.Plan del Masterclass:<ol><li>Una vista previa de cómo se pueden usar los ASTs</li><li>Un breve recorrido de cómo funcionan los compiladores</li><li>Una introducción a trabajar con un AST</li><li>El secreto para hacer miles de cambios en segundos</li><li>Cómo aplicar esta habilidad</li></ol>

Pragmática 101 sobre Herramientas JS Personalizadas para tu Proyecto

Introduction to a JavaScript tooling workshop based on personal experiences and the motivation behind creating custom tools for enhanced productivity. Speaker shares insights into managing ADHD symptoms with tools. Exploration of building tools like browser extensions and ASTs for different languages and stacks. Understanding ASTs, nodes, and types, emphasizing their significance in tools like ESLint. Creation of ESLint rules for import specifiers and identifiers differentiation. Troubleshooting ESLint rules creation process and comparing rule approaches. Leveraging AI for code automation and creating ergonomic utilities. Maximizing CodeMod efficiency and streamlining library upgrades. Implementing code refactoring with JS CodeShift and resolving AST manipulation challenges. Exploring AI tools for workflow efficiency and custom tools for codebase documentation. Evolving web development tools for accessibility and AI integration. Workshop conclusion on continuous learning, engagement, and collaboration.

Let’s demystify the secrets of modern JavaScript tooling. In this workshop, you will learn how linters, compilers, and code transforms work. You will take your first steps building your own developer tools, writing an ESLint rule, a compiler plugin, and a code transformation. Learn these skills to help your team and deepen your understanding of computer languages.Workshop Plan:<ol><li>A preview of how ASTs can be used</li><li>A brief walkthrough of how compilers work</li><li>An intro to working with an AST</li><li>The secret to making thousands of changes in seconds</li><li>How to apply this skill</li></ol>

Pragmatic 101 on Custom JS Tooling for Your Project

Las cosas están evolucionando constantemente y es difícil saber qué vale la pena. Como alguien que aprende y enseña para ganarse la vida, Scott explora la plataforma web y lo que realmente importa.

Manteniéndose al Día con el Rápido y Furioso Web

Scott Talinsky's journey in web development highlights his transition from YouTube tutorials to creating educational content and courses, reflecting his passion for teaching. The evolution of web development tools showcases advancements from limited colors to modern frameworks like React and CSS features like squircles. Adapting to new technologies emphasizes core skills and freedom in tool selection. User-centric development prioritizes accessibility and understanding the purpose of tools. Learning strategies focus on fundamentals, hands-on learning, and avoiding influencer bias. Standing out in tech roles requires quality projects, broad knowledge, and effective communication.

Things are evolving constantly and it's hard to know what's worth your time. As someone who learns and teaches for a living, Scott explores the web platform and what really matters.

Keeping Up with The Fast and Furious Web

Scott Tolinski 

JavaScript representa una gran parte del peso de la página, y optimizar para el rendimiento a menudo requiere recortar u optimizar el bundle, aunque de manera imperfecta. Al mismo tiempo, los bundlers nunca han sido más importantes en los frameworks JS de full-stack. En esta sesión, replantearemos cómo los bundlers pueden optimizar JS, si dejamos de lado algunas de nuestras suposiciones previas sobre lo que significa el bundling.

Repensando el Bundling: Por Qué No Necesitas Menos JavaScript

The speaker, Daniel, shares insights on his role as an open source maintainer, projects Nuxt and Nitro, and living in Edinburgh. Nuxt, a full-stack framework, evolved with community support, emphasizing community-driven development. Optimization strategies for JavaScript, CSS, and font loading in Nuxt were discussed. Integration of Nuxt Scripts for web performance, challenges in meta framework development, and considerations for Nuxt V3 rewrite were highlighted. Future plans include smoother migrations and open-source access to emoji slides integration.

JavaScript represents a big part of page weight - and optimising for performance often requires trimming or optimising the bundle - albeit imperfectly. At the same time, bundlers have never been more important in full-stack JS frameworks. In this session we’ll be rethinking how bundlers can optimise JS, if we let go of some of our previous assumptions about what bundling means.

Rethinking Bundling: Why You Don’t Need Less JavaScript

Daniel Roe

Aprenderás cómo ir más allá de los prompts básicos para crear flujos de trabajo integrales que proporcionen a tus asistentes de codificación de AI toda la arquitectura, las mejores prácticas, las reglas del proyecto, ejemplos de código y procesos de validación que necesita para realizar el trabajo de manera efectiva. A través de ejercicios prácticos con Claude Code, demostraremos cómo la Ingeniería de Contexto adecuada desbloquea un nivel completamente nuevo de ingeniería agente mientras enseñamos técnicas prácticas, plantillas y mejores prácticas que ayudan a los asistentes de AI a comprender los estándares de tu proyecto y entregar código listo para producción de manera consistente.

Técnicas Avanzadas de Claude Code: Ingeniería Agente con Desarrollo Impulsado por Contexto

Welcome to a guide on context engineering, focusing on using errors to our advantage and building a self-evolving AI coding assistant. Transitioning from engineering to AI development, embracing opportunities from personal automations to enterprise-grade AI deployments. Evolution of trust in AI coding assistants by leveraging errors for system evolution and self-correction. PIV loop as a main mental model for coding process, emphasizing active involvement, structured planning, and implementation. System evolution involves identifying gaps, adapting based on analysis, testing, and live demo update processes. Code validation process improvement through better validation commands and workflow optimization. Utilizing end-to-end testing for error detection and system improvement in coding assistant systems.

You'll learn how to move beyond basic prompting to create comprehensive workflows to give your AI coding assistants all the architecture, best practices, project rules, code examples, and validation processes it needs to get the job done effectively. Through hands-on exercises with Claude Code, we'll demonstrate how proper Context Engineering unlocks a whole new level of agentic engineering while teaching practical techniques, templates, and best practices that help AI assistants understand your project standards and deliver production-ready code consistently.

Advanced Claude Code Techniques: Agentic Engineering With Context Driven Development

Cole Medin

Los Signals y la reactividad de grano fino se han convertido recientemente en elementos básicos en los frameworks modernos de JavaScript. Sin embargo, hace 9 años, cuando Ryan Carniato creó SolidJS, el plano para el framework moderno, ese no era el caso en absoluto. Hoy en día, más desarrolladores usan Signals que nunca antes, pero estamos solo al comienzo del viaje. Únete a Ryan mientras mira hacia el futuro de la reactividad de grano fino, el desarrollo web, y explora lo que hay más allá de los Signals.

Beyond Signals: El Próximo Gran Cambio en la Reactividad Web

Ryan Carniato, author of SolidJS, discusses signals in the front-end world and their impact on reactivity. Signals in Solid JS have revolutionized frontend UI development by bringing signals to the forefront, impacting frameworks beyond React. Optimizing React performance with fine-grain rendering allows for updating only the necessary components, eliminating unnecessary reruns and enhancing performance composition. Mutability in programming offers benefits like stable references and reduced cloning needs. Stores in frameworks like Solid.js maintain references for unchanged data, reducing overnotifications and unnecessary recalculations. Signals enable fine-grained control with push and pull mechanisms, showcasing examples in counters, async operations, and error handling within UI components.

Signals and fine-grained reactivity have recently become staples in modern JavaScript frameworks. However, 9 years ago, when Ryan Carniato created SolidJS, the blueprint for the modern framework, that was not the case at all. Today more developers use Signals than ever before, but we are only at the beginning of the journey. Join Ryan as he looks to the future of fine-grained reactivity, web development, and explores what lies beyond Signals.

Beyond Signals: The Next Big Shift in Web Reactivity

Ryan Carniato

Los frameworks de frontend están convergiendo cada vez más, no solo en características como Signals o soporte de primera clase para TypeScript, sino también en herramientas compartidas. Vite ahora impulsa la mayoría de los ecosistemas, y los meta-frameworks dependen cada vez más de Nitro y otras capas unificadas.En esta masterclass, veremos la siguiente fase de esta convergencia: el auge de las herramientas impulsadas por Rust que mejorarán nuestros flujos de trabajo. Primero está Rolldown, un empaquetador de próxima generación que ofrece no solo velocidad o compatibilidad con Rollup, sino también características avanzadas de optimización como el control granular de fragmentos. Otra herramienta, en realidad un conjunto de herramientas que es parte de la ecuación, es Oxc, el Oxidation Compiler, una cadena de herramientas modular para analizar, transformar, minificar, pero también para linting y formateo de JavaScript a gran escala.Te irás con una imagen más clara de hacia dónde se dirige el tooling de frontend y cómo puedes adoptar estas herramientas hoy para beneficiarte del rendimiento y las características avanzadas listas para usar.

¿El Fin de la Fragmentación del Frontend? Cómo Rust está Unificando Frameworks y Herramientas

Starting a Greenfield project involves choosing frameworks, testing tools, linting, formatting, and runtime with the aim of unifying frameworks and tooling using Rust. The End of Framework Fragmentation with Rust aims for a great Developer Experience (DX) by choosing frameworks and tools based on ease of use, integration, documentation, ecosystem, speed, reliability, and future-proofing. Development benefits from the Vite server with HMR, native ESM, and optimized production builds, aiming to unify tools into the Rolldown bundler. Understanding dependency graphs, build steps, JSX, TypeScript, and bundle optimization are crucial for code organization and user experience enhancement. Bundling optimizes code for faster load times, while tool selection streamlining simplifies production builds. Vappack and OXE offer optimization features, and module resolution, code transformation, build transformation, and minification optimization play key roles in performance. Benchmarking shows SWC and OXC Minify performance comparison, and linting and formatting tools like OxLint and OXC Minify enhance development efficiency. VTest provides seamless testing, aiming for performance and compatibility, with plans for roll on Vite to become the default bundler for smoother processes.

Frontend frameworks are increasingly converging, not just in features like Signals or first class TypeScript support, but also in shared tooling. Vite now powers most ecosystems, and meta-frameworks increasingly rely on Nitro and other unified layers.In this talk, we’ll look at the next phase of this convergence: the rise of Rust-powered tooling that will improve our workflows. Up first is Rolldown, a next-gen bundler offering not only speed or compatibility with Rollup, but also advanced optimization features like granular chunk control. Another tool, actually a set of tools that is part of the equation is Oxc, the Oxidation Compiler, a modular toolchain for parsing, transforming, minifying, but also linting and formatting JavaScript - at scale.You’ll leave with a clearer picture of where frontend tooling is heading and how you can adopt these tools today to benefit of the out-of-the-box performance and advanced features.

The End of Frontend Fragmentation? How Rust is Unifying Frameworks & Tooling

El diseño a código avanza más rápido cuando tu agente tiene el contexto adecuado en cada paso. Mostraré cómo el Model Context Protocol (MCP) te permite unir servidores para que el agente pueda extraer señales de diseño estructuradas de Figma – variables, componentes, notas de diseño y mapeos de Code Connect – y emparejarlas con las reglas de tu editor/proyecto para tomar decisiones de implementación fundamentadas. A partir de ahí, agregar algunos servidores más para documentación, verificaciones en tiempo de ejecución en un navegador real y control de versiones le da al agente entradas más ricas y verificables en lugar de conjeturas. El resultado: menos desvíos "rebeldes" porque las decisiones están ancladas a fuentes reales y menos cambios de contexto ya que los desarrolladores permanecen en el IDE mientras el agente recopila lo que necesita. Es un camino práctico desde la intención de diseño hasta los cambios listos para producción, impulsado por un contexto componible en lugar de integraciones únicas.

Fast Track Design to Code With Figma MCP

Riccardo, a developer advocate at Figma, discusses the challenges of transitioning from design to code and the role of a developer advocate in bridging the gap between designers and developers. Figma's tools like DevMod, Code Connect, and Figma MCP server aim to bridge the gap by providing seamless connections between design and code. The Figma MCP server enhances design-to-code workflow by improving agent understanding of design elements, resulting in better code generation and alignment with design intent. Recommendations for optimizing design-to-code output include reusing components, implementing Code Connect, connecting design variables to code tokens, naming layers appropriately, utilizing AutoLayout, and adding annotations for clarity and context in the code. Building and sharing a design system involves generating iOS UI code with specific language and framework, and open-sourcing a design system with custom rules and guidelines for agents. Enhancing agent capabilities through multiple MCP server connections for efficient design-to-code synchronization and communication is crucial.

Design-to-code moves faster when your agent has the right context at each step. I’ll show how the Model Context Protocol (MCP) lets you stitch together servers so the agent can pull structured design signals from Figma – variables, components, design notes, and Code Connect mappings – and pair them with your editor/project rules to make grounded implementation choices. From there, adding a few more servers for documentation, runtime checks in a real browser, and version control gives the agent richer, verifiable inputs instead of guesswork. The payoff: fewer “rogue” detours because decisions are anchored to real sources and less context-switching since developers stay in the IDE while the agent gathers what it needs. It’s a practical path from design intent to production-ready changes – powered by composable context rather than one-off integrations.

Riccardo Erra

npm impulsa el ecosistema de JavaScript, pero muchos desarrolladores aún lo tratan como una caja negra.Esta charla tiene como objetivo desmitificar los flujos de trabajo principales de npm y proporcionar a los desarrolladores con varios niveles de experiencia un modelo mental más claro de cómo funciona npm detrás de escena. Con el auge de la codificación asistida por IA, muchos ingenieros están enviando código que depende de npm sin entenderlo profundamente. Mientras tanto, los desarrolladores experimentados a menudo luchan con sutilezas en torno a la publicación, la resolución de dependencias y la seguridad.Esta charla aclara el ciclo de vida de instalación y publicación, presenta las mejores prácticas modernas y ofrece herramientas de toma de decisiones que ayudan a los desarrolladores a evitar sorpresas y recuperar el control sobre sus herramientas.

Desmitificando npm: Qué Sucede Realmente Cuando Instalas y Publicas

Today's presentation delves into the inner workings of npm install and npm publish, aiming to enhance understanding and productivity. The exploration of npm CLI, ideal tree construction, and dependency types sheds light on package management intricacies. Peer dependency conflicts, resolution methods, and auditing processes are crucial for package integrity. Efficiency through symlink management and lockfile versions, along with security measures like typosquatting prevention, highlight NPM's commitment to user safety. Emphasizing the importance of maintaining updated dependencies and vigilance against malicious packages, the Talk also discusses upcoming token-related changes for secure publishing.

npm powers the JavaScript ecosystem, but many developers still treat it as a black box.This talk aims to demystify core npm workflows and give developers with various levels of experience a clearer mental model of how npm works behind the scenes. With the rise of AI-assisted coding, many engineers are shipping code that depends on npm without deeply understanding it. Meanwhile, seasoned developers often struggle with subtleties around publishing, dependency resolution, and security.This talk clarifies the install and publish lifecycle, surfaces modern best practices, and offers decision-making tools that help developers avoid surprises and regain control over their tooling.

Demystifying npm: What Actually Happens When You Install and Publish

Acompáñame en un profundo análisis técnico sobre TypeGPU que, entre otras cosas, permite que tu shader de WebGPU esté compuesto de JavaScript ya sea parcial o completamente. Esto significa que una biblioteca puede estar escrita completamente en WebGPU Shading Language, pero aceptar un callback de JS que se compila e inyecta en el código del shader. Mejorando las APIs al ir más allá de solo ajustar perillas y parámetros.En mi charla, abordaré cómo descargamos la mayor parte de este trabajo a un paso de construcción mientras seguimos apoyando escenarios dinámicos, cómo mitigamos las diferencias entre JavaScript y WebGPU Shading Language, y cómo aprovechamos las herramientas existentes como el servidor de lenguaje de TypeScript para conectar sin problemas el código de CPU y GPU a través de tipos.

TypeScript en el GPU – Cerrando la Brecha Entre Reinos

Ivo Plaza introduces TypeGPU, a TypeScript toolkit for WebGPU, with a live Jelly Slider demo. The demo showcases interactive features and a customizable API for dynamic color changes. TypeGPU offers extensibility for library creators, allowing intricate customization and per-pixel operations. It demonstrates GPU material animation control, WebGPU support, efficient debugging with GPU console log, and memory allocation. WebGPU enables program execution, resembling REST API interactions between front-end and back-end. TypeGPU integrates JavaScript shaders with WebGPU shading language, utilizing TGPU Resolve API for shader conversion and shader code generation. It covers compile-time operations, function calls, runtime conditions in shaders, and code optimization. TypeGPU's TINYEST format, JS to AST conversion, WebGPU usage, and integration with 3JS and React 3 Fiber are discussed, highlighting the differences between WebGL and WebGPU. The Talk emphasizes the integration of TypeGPU with 3JS and React 3 Fiber, utilizing GPU directives, marketing effects, AI for shader conversion, and future possibilities.

Join me for a technical deep-dive on TypeGPU which, among other things, allows your WebGPU shader to be made up of JavaScript either partially or entirely. This means a library can be written fully in WebGPU Shading Language, yet accept a JS callback that gets compiled and injected into the shader code. Improving APIs by going past just tweaking knobs and parameters.In my talk, I’ll be touching upon how we offload most of this work to a build-step while still supporting dynamic scenarios, how we mitigate the differences between JavaScript and WebGPU Shading Language, and how we leverage existing tooling like the TypeScript language server to make CPU and GPU code connect seamlessly through types.

TypeScript on the GPU – Bridging the Gap Between Realms

Iwo Plaza

Hoy en día es posible crear sistemas avanzados de seguimiento ocular sin hardware especializado, simplemente usando una cámara web estándar y la inevitable AI. Pero la verdadera diferencia la hace el modelo que ya está completamente en el edge y, por lo tanto, en el dispositivo o navegador, ¡permitiéndote trabajar incluso sin conexión! Con tecnologías de 'AI on the edge' para la detección facial y ocular, un poco de regresión lineal y un sistema de calibración en tiempo real que se adapta al usuario y al entorno, mostraré, usando solo JavaScript y modelos de AI en el dispositivo, los resultados del experimento, discutiendo la precisión, las limitaciones actuales y las posibles mejoras. Los desafíos enfrentados en la optimización en el dispositivo son muchos para equilibrar precisión y rendimiento, pero este experimento abre nuevas posibilidades para aplicaciones accesibles en la interacción humano-máquina, estudios de comportamiento y sistemas de accesibilidad, ¡todo creado con tecnologías de código abierto y hardware común!

Eye Tracking & a11y on the Edge: ¡Construye un Seguimiento Ocular con AI Ejecutándose en tu Dispositivo!

The speaker delves into the integration of AI and eye tracking technology to enhance user interactions and experiences, emphasizing the need for cost-effective solutions with standard hardware. The advancement of AI models for local device execution is highlighted, enabling cross-platform data privacy and offline usage. MediaPipe Solutions, particularly Face Landmark Detection, offer detailed facial information in real-time, and iris feature extraction is utilized for precise gaze prediction. Model calibration and mapping techniques are discussed for improving gaze interaction accuracy and reliability.

Today it's possible to create advanced eye tracking systems without specialized hardware but simply using a standard webcam and the inevitable AI. But the real difference is made by the model that is already completely on the edge and therefore on the device or browser, allowing you to work even offline! With 'AI on the edge' technologies for facial and eye detection, a bit of linear regression, and a real-time calibration system that adapts to the user and environment, I will show, using only JavaScript and on-device AI models,  the results of the experiment, discussing accuracy, current limitations, and potential improvements. The challenges faced in on-device optimization are many to balance precision and performance, but this experiment opens new possibilities for accessible applications in human-machine interaction, behavioral studies, and accessibility systems, all created with open source technologies and common hardware!

Eye Tracking & a11y on the Edge: Build an Eye Tracking With AI Running on Your Device!

Node.js continúa prosperando a pesar de las afirmaciones de su declive. Desmentiremos mitos, destacaremos mejoras recientes y demostraremos su sólido rendimiento en el panorama tecnológico actual. La charla cubrirá el papel de Node.js en la programación del lado del servidor y aplicaciones nativas de la nube, contribuciones de la comunidad, integración de características modernas de JavaScript e influencia de tecnologías emergentes. Concluiremos proyectando el brillante futuro de Node.js y áreas de crecimiento en el mundo tecnológico en evolución.

El Estado de Node.js 2025

Matteo Collina presents an updated talk on Node.js, discussing roles, open-source projects, and dispelling myths. He highlights Node.js vitality, contrasts with legacy technologies like COBOL and jQuery. The talk covers Node.js popularity, software reuse, module growth, and download stats. It addresses risks of not updating Node.js, LTS support phases, version updates, adoption rates, security measures, Node 24 features like ESM support and stable permissions. Collaboration processes, TSC role, governance, and benefits of Platformatic VAT for app development are also discussed.

Node.js continues to thrive despite claims of its decline. We'll debunk myths, highlight recent enhancements, and demonstrate its robust performance in today's tech landscape. The talk will cover Node.js's role in server-side programming and cloud-native applications, community contributions, integration of modern JavaScript features, and influence of emerging technologies. We'll conclude by projecting Node.js's bright future and growth areas in the evolving tech world.

The State of Node.js 2025

Matteo Collina

Kubernetes no tiene que ser difícil. En esta charla, exploraremos kubernetes-fluent-client, un meta framework basado en TypeScript que hace que trabajar con la API de Kubernetes sea intuitivo incluso si nunca has tocado client-go. Desde construir controladores hasta gestionar subrecursos como logs, scale y proxy, esta charla profundiza en cómo el cliente fluido simplifica la interacción con el clúster a través de una API clara y expresiva, server-side apply e informadores personalizados. Ya sea que estés escribiendo automatización u operadores completos, saldrás listo para construir aplicaciones nativas de Kubernetes más rápido y con menos código repetitivo. KFC es el Kubernetes Client que sirve como motor para <a href="https://github.com/defenseunicorns/pepr" target="_blank">Pepr</a> (Un Admission Controller y Operator Framework construido en TypeScript) y es un paquete npm.

Meta Framework para Kubernetes: TypeScript Meets Cluster Control

As TypeScript and JavaScript enthusiasts, the Talk introduces Kubernetes, Pepper, and admission control with humor and passion. Topics cover Kubernetes security, network policies, automation with Pepper, developer productivity enhancement, troubleshooting TypeScript informer, Go implementations, and comparing efficiency with DRPC streaming. The discussion extends to Kubernetes management features, community engagement, beginners' guidance, small application development goals, and auto-scaling strategies for Kubernetes clusters.

Kubernetes doesn’t have to be hard. In this talk, we’ll explore kubernetes-fluent-client, a TypeScript-based meta framework that makes working with the Kubernetes API intuitive even if you’ve never touched client-go. From building controllers to managing subresources like logs, scale, and proxy, this talk dives into how the fluent client simplifies cluster interaction through a clear, expressive API, server-side apply, and custom informers. Whether you’re writing automation or full-blown operators, you’ll walk away ready to build Kubernetes-native applications faster and with less boilerplate. KFC is the Kubernetes Client that serves as the engine for <a href="https://github.com/defenseunicorns/pepr" target="_blank">Pepr</a> (An Admission Controller and Operator Framework built in TypeScript) and is an npm package.

Meta Framework for Kubernetes: TypeScript Meets Cluster Control

Casey Wylie

Descubre cómo JavaScript y canvas sketch transforman el código mundano en arte visual fascinante en esta reveladora sesión. Esta charla revela técnicas prácticas para generar obras maestras algorítmicas utilizando principios matemáticos simples, demuestra ejemplos de codificación en vivo que difuminan la línea entre el desarrollo y el arte, y comparte cómo este enfoque de codificación creativa revolucionó mi propio trabajo. Ya seas un desarrollador experimentado o un codificador creativo, te irás con las habilidades para expresar tu visión artística a través del código y unirte a la creciente comunidad de artistas computacionales que están empujando los límites de la creación digital.

Dijeron que JavaScript no podía hacer arte hermoso – Estaban equivocados

The Talk delves into JavaScript artwork creation, debunking misconceptions and exploring its evolution from basic animation to advanced web development innovations. It discusses the development of animation technologies in JavaScript, comparing Canvas and WebGL, and emphasizes the diversity of tools available. The session focuses on using Canvas Sketch for creating custom and abstract designs, detailing rendering, saving points, working with coordinates and styles, and utilizing animation for interactivity. It also covers specifying animation parameters and exploring options for connectivity.

Discover how JavaScript and canvas sketch transform mundane code into mesmerizing visual art in this eye-opening session. This talk reveals practical techniques to generate algorithmic masterpieces using simple math principles, demonstrates live coding examples that blur the line between development and artistry, and shares how this creative coding approach revolutionized my own work. Whether you're a seasoned developer or creative coder, you'll walk away with the skills to express your artistic vision through code and join the growing community of computational artists pushing the boundaries of digital creation.

They Said JavaScript Couldn't Make Beautiful Art – They Were Wrong

Raihan Nismara

En este masterclass, guiaré a los participantes a través de un proceso para construir aplicaciones full-stack usando Cursor. Cubriremos la configuración de flujos de trabajo de Cursor, incluyendo reglas, comandos y modos personalizados, para agilizar el desarrollo. Usaremos Cursor para redactar PRDs, esquemas de base de datos, etc., y un SOW simple. Convertiremos notas en listas de verificación accionables para guiar la construcción, conectando el front end, API y base de datos.Además, crearemos flujos de trabajo para depurar, probar, revisar y corregir la aplicación mientras reducimos alucinaciones con indicaciones precisas. La sesión también incluirá un flujo de trabajo de revisión de seguridad y técnicas para aplicar indicaciones para lograr un diseño de UI moderno. Los asistentes se irán con una aplicación funcional en su máquina y un proceso de desarrollo reutilizable.

Cómo Construir Aplicaciones Full-Stack Usando Cursor

Ray conducting a workshop on building full stack apps using Cursor, emphasizing the importance of rules and demonstrating rule creation. Exploring advanced rule application and tool management for efficient context handling. Focused on customized planning modes and resource sharing in Cursor. Utilizing Composer for efficient workflow iteration and encountering unexpected file creation issues. Defining detailed statements of work for app development and enabling tools like Superbase CLI and Chrome for testing. Troubleshooting bugs, creating custom commands for automation, and self-correcting agent behavior for quality assurance. Encouraging feedback loops, pull request creation, and building reusable workflows with Coderabbit review.

In this workshop, I’ll guide participants through a process for building full-stack applications using Cursor. We’ll cover setting up Cursor workflows, including rules, commands, and custom modes, to streamline development. We will use Cursor to draft PRDs, database schema, etc and a simple SOW. We’ll turn notes into actionable checklists to guide the build, connecting the front end, API, and database.Additionally, we’ll create workflows for debugging, testing, reviewing, and fixing the app while reducing hallucinations with precise prompts. The session will also include a security review workflow and techniques for applying prompts to achieve a modern UI design. Attendees will leave with a functioning app on their machine and a reusable development process.

How to Build Full-Stack Apps Using Cursor

 Shrey Shah

El manejo de archivos en JavaScript va mucho más allá de fs.readFile. En esta charla práctica, aprenderás cómo procesar archivos grandes de manera eficiente utilizando streams, pipeline(), iteradores asíncronos y manejo de backpressure, todo sin colapsar tu servidor. Desde el análisis en tiempo real de CSV hasta la descompresión de ZIP y la limpieza de archivos segura para producción, exploraremos patrones poderosos que hacen que tus aplicaciones de Node.js sean más rápidas, seguras y listas para escalar.

Manejo de Archivos en JavaScript Como un Profesional: Desde ZIP Streams hasta Análisis Eficiente de Memoria

Node.js offers significant capabilities for efficient file handling, emphasizing the importance of processing files effectively to avoid crashes and memory issues. Understanding streams, backpressure, and stream concepts like readable, writable, duplex, and transform streams is crucial for efficient file handling. Streams enable working with data in chunks, ensuring stable memory usage and efficient applications. Backpressure in streams optimizes memory usage by processing data incrementally, unlike loading entire files into memory. Utilizing readable streams in processing improves concurrency and avoids blocking the event loop.

Handling files in JavaScript goes far beyond fs.readFile. In this practical talk, you'll learn how to process large files efficiently using streams, pipeline(), async iterators, and backpressure handling — all without crashing your server. From real-time CSV parsing to ZIP decompression and production-safe file cleanup, we’ll explore powerful patterns that make your Node.js apps faster, safer, and ready for scale.

JavaScript File Handling Like a Pro: From ZIP Streams to Memory-Efficient Parsing

La plataforma web está evolucionando, introduciendo nuevos primitivos poderosos que cambian la forma en que construimos interfaces. Entre estas innovaciones se encuentra el toplayer, un concepto de renderizado importante pero a menudo malentendido. Esta charla explora el mecanismo de renderizado del toplayer, aprendiendo cómo habilita patrones de UI y examinando su papel en la accesibilidad, el rendimiento y la experiencia del desarrollador.El toplayer no es solo otra especificación técnica: representa un cambio en cómo los navegadores manejan elementos que necesitan salir del flujo normal del documento. Desde elementos de diálogo hasta la nueva API de popover, tooltips personalizados hasta interfaces modales, el toplayer se ha convertido silenciosamente en esencial para crear experiencias web modernas. Al comprender cómo funciona el toplayer, los desarrolladores pueden construir componentes más robustos, accesibles y mantenibles, evitando errores comunes que llevan a experiencias de usuario inconsistentes.En esta sesión, profundizaremos en el toplayer para entender cómo los navegadores manejan contextos de apilamiento, orden de pintura y gestión de enfoque dentro del toplayer. A través de ejemplos prácticos y demostraciones de codificación en vivo, los asistentes obtendrán tanto comprensión teórica como habilidades prácticas que pueden aplicar a sus propios proyectos.

Sumergiéndose en el Toplayer: Donde Viven los Diálogos, Popovers y Modales

Online talk discussing web overlays, ads, models, and cookie banners. Focus on dialogues, popovers, and their history in web design. Emphasis on accessibility and successful UI strategies. Introduction of non-modal dialogues and popovers in web design. Insights on implementing model dialogues, popovers, and managing Z-index values. Exploration of Z-index challenges, top layer concepts, and dialogue accessibility. Analysis of dialogue opening, focus management, and urgent model dialogues. Implementation of advanced popover features, CSS anchor usage, and dialogue comparisons.

The web platform is evolving, introducing powerful new primitives that change how we build interfaces. Among these innovations is the toplayer – an important yet often misunderstood rendering concept. This talk explores the toplayer rendering mechanism, learning how it enables UI patterns and examining its role in accessibility, performance, and developer experience.The toplayer isn't just another technical specification—it represents a shift in how browsers handle elements that need to break out of the normal document flow. From dialog elements to the new popover API, custom tooltips to modal interfaces, the toplayer has quietly become essential for creating modern web experiences. By understanding how the toplayer works, developers can build more robust, accessible, and maintainable components while avoiding common pitfalls that lead to inconsistent user experiences.In this session, we'll dive into the toplayer to understand how browsers handle stacking contexts, painting order, and focus management within the toplayer. Through practical examples and live coding demonstrations, attendees will gain both theoretical understanding and practical skills they can apply to their own projects.

Diving Into the Toplayer: Where Dialogs, Popovers, and Modals Live

Tim Damen

Los compiladores han sido vistos durante mucho tiempo como uno de los temas más complejos en la ciencia de la computación. Hoy en día, los frameworks web están evolucionando de bibliotecas de tiempo de ejecución a compiladores de optimización y remodelando cómo construimos interfaces de usuario. Diferentes temas en el desarrollo front-end—por ejemplo, la programación reactiva—están experimentando ahora esta era impulsada por conocimientos basados en análisis estático.Esta charla explorará cómo los compiladores modernos de UI están redefiniendo el rendimiento y la experiencia del desarrollador. Discutiremos cómo se diseñan algunos compiladores y diferentes soluciones de código abierto — incluyendo el React Compiler, Million.js, Svelte, y Marko, así como aplicaciones empresariales, y el creciente papel de los compiladores en la automatización de tareas.¡Al final de la sesión, probablemente serás optimista sobre un futuro donde los compiladores entiendan todo nuestro código y descarguen una gran parte de nuestra carga mental relacionada con tareas manuales!

Compiladores, Interfaces de Usuario y el Resto de Nosotros

Speaker shared a story of creating a Haskell-inspired language for Zebra labelers using a compiler. Discussed the evolution of JS frameworks into compilers, focusing on React compiler's advanced techniques. Highlighted the role of compilers in reactive programming and code optimization. Explored challenges in large-scale migration using tools like tsmorph and QuickType. Emphasized the significance of compilers in user interface development and software architecture decisions. Addressed the future of front-end tools, Web Assembly adoption, and the importance of experimentation with tools for project success.

Compilers have long been seen as one of the most complex topics in computer science. Nowadays, web frameworks are evolving from runtime libraries into optimizing compilers and reshaping how we build user interfaces. Different topics in front-end development—e.g., reactive programming—are now experiencing this era fueled by static-analysis-driven insights.This talk will explore how modern UI compilers are redefining performance and developer experience. We’ll discuss how some compilers are designed and different open-source solutions — including the React Compiler, Million.js, Svelte, and Marko, as well as enterprise apps, and the growing role of compilers in automating things.By the end of the session, you'll probably be optimistic about a future where compilers understand our entire code and offload a huge part of our mental burden related to manual tasks!

Compilers, User Interfaces & the Rest of Us

Matheus Albuquerque

Electron conecta las aplicaciones web y de escritorio, permitiendo a los desarrolladores convertir aplicaciones web en experiencias de escritorio multiplataforma. Pero más allá del inicio rápido, se encuentra una serie de desafíos de ingeniería, desde dolores de cabeza con la firma de código hasta cuellos de botella en CI/CD y riesgos de seguridad.

Construyendo Aplicaciones de Escritorio a Partir de Aplicaciones Web: Escalando Electron con Firma de Código y CI/CD

Ayole Aonseola discusses converting web apps into desktop applications using Electron, emphasizing challenges with code signing, CI CD, and industrial scaling. Electron architecture focuses on security with context isolation and node integration. Engineering hurdles include distributing applications across OSs, implementing automation for error prevention, and ensuring app authenticity through code signing. Mac-specific configurations, like app bundle ID and notarization, are crucial. Building universal apps for Mac and Windows requires understanding differences in building environments and certificate access. Automation of signing processes with Azure and GitHub streamlines workflows, handling EV tokens and securely managing certificates. CICD automation includes base64 certificate encoding and triggering build processes on tag pushes.

Electron bridges the web and desktop apps, enabling developers to turn web apps into cross-platform desktop experiences. But beyond the quick start lies a series of engineering challenges, from code signing headaches to CI/CD bottlenecks, and security risks.

Building Desktop Apps Out of Web Apps: Scaling Electron With Code Signing & CI/CD

Ayodele Aransiola

Los flujos de trabajo duraderos - flujos de trabajo que registran su estado para recuperarse automáticamente de fallos - permiten a los desarrolladores construir código confiable más rápido y reducir drásticamente la gravedad de los incidentes de producción. La mayoría de los sistemas de flujo de trabajo requieren que configures un montón de infraestructura, ¡pero eso no es necesario! En esta charla, te mostraremos cómo construir flujos de trabajo duraderos en puro JavaScript, como una biblioteca que cualquier aplicación puede importar.

Construyendo Flujos de Trabajo Duraderos Desde Cero en JavaScript

Peter discusses building durable workflows in pure JavaScript, highlighting challenges in creating reliable systems, especially in complex applications like Money Transfer. Failures can occur due to various reasons, including process crashes, resource issues, timeouts, and API failures. Durable workflows involve checkpointing for recovery, traditionally involving heavyweight external orchestration systems. Designing a JavaScript workflow library aims for simplicity and durability, enabling checkpointing directly on app servers. Implementing the library includes checkpoints for safety, recovery, and workflow integrity maintenance.

Durable workflows - workflows that checkpoint their state to automatically recover from failure - enable developers to build reliable code faster and dramatically reduce the severity of production incidents. Most workflow systems require you to set up a bunch of infrastructure, but that’s not necessary! In this talk, we’ll show you how to build durable workflows in pure JavaScript, as a library any application can import.

Building Durable Workflows From Scratch in JavaScript

Peter Kraft

Probablemente te has enfrentado a una situación donde necesitabas depurar algo particular en tu aplicación. Algo en una condición específica. Este tipo de situaciones tienden a ser mal manejadas por nosotros los desarrolladores.Un enfoque común es agregar un montón de logs, desplegar el código y esperar a que los logs regresen. Este es un proceso muy lento y puede ser muy frustrante.Sería mucho mejor si pudieras simplemente entrar en el entorno, mientras se ejecuta, ejecutar algún código y ver los resultados. Aquí es donde entra el módulo node:repl. Te permite crear un Read-Eval-Print Loop (REPL) en tu aplicación Node.js.Inspirado por Ruby on Rails, que tiene una consola muy útil, implementé una consola en la aplicación en la que estaba trabajando. Fue un cambio radical. Ahora, tenemos un proceso de depuración/experimentación mucho más efectivo. Mi equipo se enamoró de él. Ahora, es parte de la aplicación, como si siempre lo hubiera sido.Otra cosa buena es que no fue difícil de implementar. El módulo node:repl lo da todo. Solo necesitamos asegurarnos de que la aplicación esté correctamente inicializada para usar el módulo REPL.Tabla de contenidos<ul><li>La importancia de las herramientas y DX</li><li>El módulo node:repl</li><li>Práctica: creando un REPL iterativo para depurar cualquier aplicación</li><li>Conclusión</li></ul>Conclusiones<ul><li>Por qué DX importa y cómo mejorarlo a través del desarrollo de herramientas personalizadas</li><li>Cómo los módulos de Node.js pueden ayudar con la creación de herramientas efectivas</li><li>Conocer el módulo node:repl</li></ul>

Depuración interactiva y control en Node.js con REPL

Eddie introduces Interactive Debugging and Control in Node.js with Rapple, addressing debugging challenges in production environments. Node.js lacks a universal tool like Rails for interactive control. Challenges in bootstrapping Node.js applications include diverse tool selections and careful setup. Developing a custom console with Node-REPL improves team efficiency. Implementing a console with database manipulation using Objection ORM enhances functionality. Testing and sandbox mode implementation for interactive debugging in Node.js are explored. Adonis framework benefits include ORM and Ace REPL for seamless development and experimentation.

You probably have faced a situation where you needed to debug something particular in your application. Something in a specific condition. These kinds of situations tend to be poorly handled by us developers.A common approach is adding a bunch of logs, deploying the code, and waiting for the logs to return. This is a very slow process and can be very frustrating.It would be much better if you could just get into the environment, while running, execute some code, and see the results. This is where the node:repl module comes in. It allows you to create a Read-Eval-Print Loop (REPL) in your Node.js application.Inspired by Ruby on Rails, which has a very useful console, I implemented a console in the application I was working on. It was a game changer. Now, we have a way more effective debugging/experimentation process. My team fell in love with it. Now, it's part of the app, as it always has been.Another good thing is that it was not hard to implement. The node:repl module gives everything. We just need to ensure the application is properly bootstrapped to use the REPL module.Table of contents<ul><li>The importance of tools and DX</li><li>The module node:repl</li><li>Hands-on: creating an iterative repl for debugging any application</li><li>Conclusion</li></ul>Takeaways<ul><li>Why DX matters and how to improve it through the development of custom tools</li><li>How Node.js modules can help with the creation of effective tools</li><li>Get to know the node:repl module</li></ul>

Interactive Debugging and Control in Node.js With REPL

Edy Silva

Imagina la conveniencia de ejecutar archivos TypeScript directamente con Node.js usando `node file.ts`. Hace solo unos años, este concepto parecía un sueño lejano. Hoy, se presenta como una característica experimental emocionante. Esta narrativa se adentra en el viaje de transformar este sueño en realidad. Integrar TypeScript de manera nativa en Node.js planteó desafíos significativos. Estos obstáculos incluyeron gestionar la falta de garantías de versionado semántico estricto de TypeScript y acomodar varios sabores de TS. Hoy, el soporte nativo experimental de TypeScript en Node.js marca un hito transformador. Más allá de simplemente ejecutar archivos .ts, significa un avance en el desarrollo de JavaScript.

Node.js + TypeScript: Una integración nativa años en desarrollo

Welcome to JS Nation. Marco Ippolito discusses NodeJS and TypeScript integration, challenges, and growth. TypeStripping in Node.js removes TypeScript syntax for execution. Efficient TypeStripping in Node 22 skips SourceMaps, aligns with TypeScript versions. Enhanced features in Node 22 support experimental TypeScript features with limitations. Node.js supports TypeScript features with erasable syntax and namespace in Node 23.8. Importing types correctly crucial to avoid errors. Node.js TypeScript integration update includes switching to TypeScript, browser support, and community involvement.

Imagine the convenience of executing TypeScript files directly with Node.js using `node file.ts`. Just a few years ago, this concept seemed like a distant dream. Today, it stands as an exciting experimental feature. This narrative dives into the journey of transforming this dream into a reality. Integrating TypeScript natively into Node.js posed significant challenges. These hurdles included managing TypeScript’s lack of strict semantic versioning guarantees and accommodating various TS flavors. Today, the experimental native TypeScript support in Node.js marks a transformative milestone. Beyond merely executing .ts files, it signifies a leap forward in JavaScript development.

Node.js + TypeScript: A Native Integration Years in the Making

Marco Ippolito

¿Hemos estado atrapados en los mismos bucles de dolor de pruebas durante años? ¡Aunque pueda parecerlo, no lo hemos estado!Incluso si AI ahora está escribiendo nuestras pruebas, muchos desarrolladores de JavaScript han enfrentado numerosos desafíos, desde la falta de claridad en las clasificaciones de pruebas hasta la lucha por el mocking, pruebas de integración de larga duración y el cambio de todo a pruebas E2E.Recapitulemos todo lo que ha sucedido en los últimos años y miremos al presente para explorar lo que 2025 traerá para las pruebas en el mundo de JavaScript en torno a los runners de pruebas, bibliotecas de pruebas, mocking, prácticas de producción y herramientas basadas en AI.

El estado de las pruebas de JavaScript en 2025

Speaker reflects on the importance of testing, shares insights from the state of JS survey, and discusses future plans for JavaScript testing in 2025. Companies are exploring testing tools and methodologies based on surveys and personal data. Evolution of testing practices from browser to node-based environment with VTest adoption. Advancements in testing technology include VTest's role, Playwright for end-to-end testing, and comprehensive testing solutions. Progress in self-healing tests, AI integration, and tool resurgence like Storybook and MSW for API mocking. Diverse opinions on testing in 2025, with a focus on static analysis, VTest, and end-to-end testing resurgence to browser.

Have we been stuck in the same testing pain loops for years? Even though it might seem like it, we haven't!Even if AI is now writing our tests, many JavaScript developers have faced numerous challenges, from the unclarity of test classifications to the struggle of mocking, long-running integration tests, and shifting everything to E2E tests.Let's recap everything that has happened in the past years and look into the present to explore what 2025 will bring for testing in the JavaScript world around test runners, testing libraries, mocking, production practices, and AI-based tools.

The 2025 State of JavaScript Testing

Daniel Afonso

Las aplicaciones modernas enfrentan amenazas de seguridad significativas, especialmente para aplicaciones JavaScript basadas en navegador. Esta breve charla te ayudará a entender cómo la arquitectura de la aplicación y la seguridad de autenticación se impactan mutuamente. Aprenderás el patrón de arquitectura más seguro y basado en estándares para implementar autenticación en aplicaciones JS.

Backend-for-Frontend Auth: La Arquitectura Segura de Aplicaciones JS

Challenges in OAuth, importance of back end security for front end, and implementation details of OAuth flow between front end, back end, and authorization server. Detailed process of token handling, session management, and user data retrieval in OAuth implementation between front end, back end, and authorization server.

Modern apps face significant security threats, especially for browser-based JavaScript apps. This short talk will help you understand how app architecture and auth security impact each other. You will learn the most secure, standards-based architecture pattern for implementing auth for JS apps.

Backend-for-Frontend Auth: The Secure JS App Architecture

Kim Maida

"Entropy Tolerance" es una idea original para ayudar a los líderes técnicos a navegar las decisiones de adopción de AI. Este concepto mide cuánta incertidumbre generada por AI pueden manejar tus sistemas y procesos antes de que se rompan. Aprende este marco de toma de decisiones para guiar a tus equipos sobre cuándo adoptar AI y cuándo proceder con precaución.

Entropy Tolerance: La Pregunta de Software Más Importante Que No Estás Haciendo

In the Talk, the focus is on understanding software determinism and the implications for AI implementation. It delves into the importance of trustworthiness, reliability, and consistency in deterministic systems. The discussion also explores non-deterministic behavior in AI, emphasizing the pros and cons of unpredictability. The significance of measuring entropy tolerance in AI integration and business processes is highlighted, along with the value assessment based on tolerance levels. Additionally, the Talk touches on hybrid collaboration in AI utilization, showcasing examples of low, medium, and high entropy tolerance scenarios, and the role of AI in security auditing and high-tolerance tasks.

"Entropy Tolerance" is an original idea to help technical leaders navigate AI adoption decisions. This concept measures how much AI-generated uncertainty your systems and processes can handle before they break. Learn this decision-making framework to guide your teams on when to embrace AI and when to proceed with caution.

Entropy Tolerance: The Most Important Software Question You Aren't Asking

Tony Alicea

Esta charla comparte lecciones del mundo real de la construcción y operación de un servidor MCP (Model Context Protocol) remoto. Cubre las decisiones arquitectónicas, los desafíos técnicos y las ideas prácticas que surgieron al habilitar la ejecución de modelos escalables y conscientes del contexto en una de las plataformas de pago más grandes del mundo.

Lecciones de la Construcción de un Servidor MCP Remoto Empresarial

Hemant, machine learning manager at PayPal, explains enabling models to access real-time data using tools and function calls. Discussion on the launch of MCP by Antropic and its attributes for standardized protocol. Overview of PayPal's MCP architecture, security measures, and best practices. Importance of trust boundaries, multi-layer security, and authentication in tool access. Evolution of tool architecture, deployment strategy, and blockchain payment protocols. LLM code execution for efficient MCP handling and enterprise-scale launch strategies. Collaboration in tech projects, future of financial MCPs, and use cases in consumer banking.

This talk shares real-world lessons from building and operating remote MCP (Model Context Protocol) server. It covers the architectural decisions, technical challenges, and practical insights that came from enabling scalable, context-aware model execution in one of the world’s largest payment platforms.

Lessons from Building Enterprise Remote MCP Server?

Hemanth HM

Durante años, hemos dependido de servidores centralizados y, más recientemente, nos hemos apoyado en gran medida en componentes del lado del servidor para impulsar nuestras aplicaciones web. Pero esta dependencia tiene un costo: aplicaciones que se rompen sin conectividad constante a Internet, rendimiento lento bajo carga pesada y frustración del usuario cuando las cosas simplemente no funcionan. Esta charla desafía el status quo de la nube primero e invita a los desarrolladores a explorar el poder de **local-first**. Imagina construir experiencias multijugador y colaborativas que prosperen sin conexión, se sincronicen sin problemas cuando estén en línea y aprovechen el valor del estado localizado. También abordaremos las complejidades de la resolución de conflictos al pasar de una autoridad centralizada a un servidor de sincronización como mediador ligero. Es hora de repensar la colaboración.

Local First – Colaboración Más Allá de la Nube

Rethinking collaborative app development for offline and online usage, introducing LocalFirst paradigm. Challenges of offline usage in collaborative app development and potential solutions. Benefits of the local first paradigm for user data control and offline work in collaborative apps. Exploring the local first app demo with Yjs technology for collaborative features and offline functionality. Exploring Yjs for collaborative data management and synchronization in React components, enabling a shift to a local first approach. Exploring the reduced role of the server in synchronization, emphasizing user ownership of data. Using Yjs shared types for familiar CRUD operations, introducing Yjs providers for real-time collaboration, and leveraging IndexedDB for offline data storage. The y-index-db-provider persists the document locally, and the y-web-socket-provider relays updates for collaborative scenarios, ensuring consistent list management. While offline, users make individual additions to a shared document. Upon reconnecting, Yjs ensures consistent ordering of concurrent offline changes, guaranteeing a synchronized state across all clients. What happens when clients edit the same item in Yjs is demonstrated. The last write wins register in Yjs ensures the last operation prevails. CRDTs handle concurrent edits and deletions, favoring removals for definitive operations. Yjs provides an elegant solution with conflict-free replicated data types.

For years, we’ve defaulted to centralised servers and, more recently, leaned heavily on server-side components to power our web apps. But this dependence has a cost: apps that break without constant internet connectivity, sluggish performance under heavy load, and user frustration when things don’t just work. This talk challenges the cloud-first status quo and invites developers to explore the power of **local-first**. Imagine building multiplayer, collaborative experiences that thrive offline, sync smoothly when online, and harness the value of localised state. We’ll also tackle the complexities of conflict resolution when shifting from a centralised authority to a sync server as a lightweight mediator. It’s time to rethink collaboration.

Local First – Collaboration Beyond Cloud

La IA empresarial es un problema de orquestación. Necesitamos sistemas que sean modulares, observables y escalables, desde nuestros centros de datos centrales hasta el edge. Esta sesión es un recorrido arquitectónico de cómo LangGraph.js proporciona el "pegamento" para conectar estos patrones empresariales modernos en un entorno TypeScript/Node.js.Comenzaremos enmarcando el agente de IA como un "Monolito Modular", un servicio único y desplegable construido a partir de componentes basados en gráficos y testeables. Esta estructura gráfica explícita no es solo para la lógica; hace de la Observabilidad un Ciudadano de Primera Clase, permitiéndonos rastrear el razonamiento cíclico del agente en producción.A continuación, desplegaremos este agente como un Backend-for-Frontend (BFF) inteligente, empoderando nuestras aplicaciones full-stack con orquestación con estado. Finalmente, abordaremos la próxima frontera: Arquitectura para el Edge. Exploraremos patrones para gestionar el estado persistente del agente en un entorno híbrido, permitiendo que nuestra IA funcione en cualquier lugar.Esta masterclass es para el arquitecto e ingeniero que necesita construir y escalar aplicaciones de IA de grado empresarial en el mundo real.

Orquestación de IA Full-Stack: Agentes Modulares, Observabilidad y el Edge

The Talk delves into artificial intelligence, deep learning, and agentic systems, emphasizing the challenges and importance of building trustworthy AI systems. It proposes a modern enterprise stack for resilient and scalable AI systems, discussing the use of modular architecture, agent logic, and observability tools. The session explores the benefits of iterative agent logic, scalability enhancements, and efficient model management. Additionally, it showcases practical applications in sports statistics verification and playoff probability estimation while highlighting the significance of system performance, efficiency optimization, and robustness in large-scale operations.

Enterprise AI is an orchestration problem. We need systems that are modular, observable, and scalable – from our central data centers to the edge. This session is an architectural tour of how LangGraph.js provides the “glue” to connect these modern enterprise patterns in a TypeScript/Node.js environment.We’ll start by framing the AI agent as a “Modular Monolith”—a single, deployable service built from testable, graph-based components. This explicit graph structure isn’t just for logic; it makes Observability a First-Class Citizen, allowing us to trace the agent’s cyclical reasoning in production.Next, we’ll deploy this agent as an intelligent Backend-for-Frontend (BFF), empowering our full-stack applications with stateful orchestration. Finally, we’ll tackle the next frontier: Architecting for the Edge. We will explore patterns for managing persistent agent state in a hybrid environment, allowing our AI to run anywhere.This talk is for the architect and engineer who needs to build and scale real-world, enterprise-grade AI applications.

JSNation US 2025

The main web dev conference in the US

Lera

Neha Seth

Brian Whippo

Simon Ouyang

Lera

Daria Shabala

Seungho Park

Will Klein

Alexander Lichter

Karen Li

Alex Astrum

Richard Carrigan

Dipanshu Gupta

Varun

Amanda Witecki

Kyle Langford

Houda Mouttalib

Pascal Moser

Trust Jamin

James O'Reilly

Mauro Bartolomeoli

Kritiketan Sharma

Joana Santos

Noam Zingboim

Henry Lie

Francesco Sciuti

Benjamin Schachter

Herrington Darkholme

Elina Denisova

David Mytton

Leon Malisov

Maya Shavin

Hunor Márton Borbély

Mangirdas Darguzis

Total attendees