Callstack

"Drunk with power" from AI coding? @matthargett.bsky.social has a question to bring you back to earth: what will your product cost to run when people use it? Watch the full React Universe On Air episode: clstk.com/4zFCUTv

"I only trust them when I don't look at the code." Tzvetan Mikov is only half joking. He also explains how he reviews specs and tests, with Codex and Claude helping with review. Watch the interview and read the show notes at clstk.com/4sicTqO

An enterprise team we worked with moved React Native adoption from under 10% to about 30% while feature delivery continued. The model: migrate screen by screen, keep stable native flows in production, and package React Native as native artifacts. Read the case study: clstk.com/4yCWsHp

An AI agent drafts an email, but needs approval to send it. David Khourshid of Stately explains how state machines define that workflow and handle unexpected model outputs. Watch the interview and read show notes at clstk.com/4sdPXJo

Checking an invitation panel means recreating the invitation every time. With Appduct, an AI agent can call a function your team writes to trigger the panel with test data, wait until it’s ready, and inspect the UI. See how: clstk.com/3VPdSCZ

Why put 100 agents on code quality? Carl Ross of OpenAI explains competing fixes, grading candidates, and continuous code cleanup as features ship. Watch the interview and read show notes at clstk.com/4sdwoAP

An agent fleet needs shared state, authentication, and observability. Nishant Gupta of Meta explains how these fit together and what it takes to trust agent outcomes. Watch the interview: clstk.com/4hZ1nNT

Tap in one Quest window and watch the other counter update. Both panels use the same parent state. The Meta VR Layout SDK keeps them in one React Native tree. Watch the demo and read more on our blog: clstk.com/4dH9ADI

AI makes it tempting to add every feature in sight. Can the product handle 1,000 users an hour? What will 10,000 cost? Matt Hargett explains why scale and quality need to be part of the agent workflow ➡️ clstk.com/4zB8TV6

Who's testing the code your agents write? @szymonrybczak.bsky.social of TesterArmy explains AI testing for web and mobile apps, including test plans built around each pull request. Watch the interview at clstk.com/3VyIPLD

Use Apex in your usual coding agent or wire it into your own tooling. The API is OpenAI- and Anthropic-compatible, so you can keep your workflow. Get API access: clstk.com/3W8dlvQ

e2e by TesterArmy runs mobile tests through our open-source agent_device. Combine goal-based agent actions with exact assertions, then replay cached actions without inference on cache hits. Read how it works on iOS and Android 📖  clstk.com/3U0Oycz

Your agents work while you sleep. So does your token bill. Join us live on October 15 📺 clstk.com/4hQnvbK Kewin and Szymon will discuss stalled runs and cost per ticket in React Native software factories. See you there!

Grok Bot reached internal beta in about a month. Nick Miller of SpaceXAI explains how Cursor's browser experience and isolated virtual machines shaped its computer use. Watch the interview: clstk.com/47MQ3y9

Have you heard that Apex is live? 🎉 It's our coding model specialized in React Native, and you can use it in Claude Code, Cursor, or Codex. Try it on a task from your project: clstk.com/4y47p4o

A crypto trading platform we worked with cut its JavaScript bundle size by about one-third. DMAIC-based performance work and CI regression checks kept improvements measurable as the React Native migration progressed. Read the case study: clstk.com/4rCS5dj

An agentic workflow needs a recovery path. We introduced an invalid commit reference, routed the review feedback back to the Developer agent, verified the correction, and then sent the task to human approval. See the case study 👉 clstk.com/4hn7Lgd

Sharing a lot of code does not automatically give you one mobile product. Share the layers whose behavior should stay consistent. Keep platform differences explicit, with owners and tests. React Native at Scale 2026 explains how experienced teams make that call: clstk.com/4xY0cmB

5x to 10x WebGPU speedups. Hugging Face is testing a browser AI inference engine using agent-written, device-optimized WebGPU kernels. Nico Martin joins Mike Grabowski to discuss the early work and what it could mean for Transformers.js. Watch React Universe On Air 🎙️ clstk.com/4xZLZpU

A React Native migration for a crypto trading platform gave coding agents project-specific skills, access to simulators and devices, and React DevTools context. This made engineering standards reusable across teams. Read the case study to learn more at clstk.com/4xgK6Uo

We routed verifiable tasks to autonomous agents and kept visual judgment in an engineer-guided Codex loop. The case study covers the workflow behind an under-one-month launch and up to 3x estimated acceleration: clstk.com/4gFcWJj

Which AI tasks belong in the browser? @nico.dev from Hugging Face covers Transformers.js, ONNX Runtime, WebGPU, model caching, fallbacks, and browser agents. See which tasks are ready for on-device use and where the limits are. Watch & read shownotes at clstk.com/4qXHf1f

Background removal and transcription run in the browser. Nico Martin from Huggin Face joins @grabbou.com to cover Transformers.js today, device limits, offline use, and which parts of an agent workflow can already run locally. 📺 YouTube premiere: youtu.be/elAVSra9KCY 📆 Sept 7 at 2PM CEST

Ruben Casas from Postman has a few words for you ahead of Agent Conf. In his talk, he’ll explore the emerging spectrum of generative UI, from static templates & configuration-driven interfaces to interfaces authored on the fly by LLMs. Get your ticket: www.agent.sh/tickets

React Native can give teams one shared codebase. Still, it does not decide who owns upgrades, releases, or the shared setup. We combined 10 years of experience with perspectives from 12 organizations in React Native at Scale 2026: clstk.com/4r1ELPt

We wrote up how we embedded React Native inside a .NET MAUI app using a brownfield setup. The article covers packaging, lifecycle wiring, MAUI Shell navigation, and shared state across both runtimes. Read the article: clstk.com/4zNHvny

Zuzanna Skoczylas from Sierra has a few words for you ahead of Agent Conf 2026. In her talk, she’ll share three lessons from building and shipping production agents: context engineering, the agent loop, and benchmarking model reliability. Tickets: www.agent.sh/tickets

Registering a native module twice can crash a React Native app. Our Super App Showcase prevents that by keeping native code in the Host Shell. Mini Apps stay JavaScript-only and load when needed with Re.Pack and Module Federation. Read more about it: clstk.com/4gsofo3

One request now starts the automated React Native Evals run. Airflow runs each solver/judge pair separately. MLflow keeps the run record and files together. Human review still decides what gets published. Here’s how we built it 👉 clstk.com/4wMcgqh

Lech Kalinowski from Callstack is joining Agent Conf in Warsaw as a speaker. Together with Artur Morys-Magiera (@artus9033.bsky.social), he’ll present a case study of Apex and look at the impact of quantization on LLM performance. Get your ticket: www.agent.sh/tickets

Jakub Kuzimski from Allegro will speak at AgentConf 2026 on Stop Making LLMs Deterministic. Building a Spec-Driven Software Factory at Allegro. Join us in Warsaw on September 17-18: www.agent.sh/agenda

We keep migration knowledge in six versioned files: 1. `MIGRATION.md`, 2. `SCREENS.tsv`, 3. `STATE_AND_STORAGE.tsv`, 4. `DEPENDENCIES.tsv`, 5. `EVENTS.tsv`, 6. `PARITY_CHECKS.md`. Teams review them in spreadsheets and Git. Agents use them for each task. Download the guide 👉 clstk.com/4quxRSF

Last night in Cambridge, around 80 people joined us for Agentic AI in the Wild with Qodo Dev Community, ReactVision, and Codex Community Boston. Talks covered code quality, local validation, spatial computing, and large codebase migration. Thanks to everyone who came!

Give agents two migration tasks: 1. Preserve native behavior 2. Refactor into idiomatic React Native A different agent reviews between passes. The team compares both apps on devices and repeats the checks after refactoring. Download the guide 👉 clstk.com/3So6LzM

Why do camera features break even on devices that technically support them? Marc Rousavy breaks down a problem every team runs into once camera work gets more advanced: HDR, 5K & 120 FPS may all be supported, but that does not mean those settings will work together. Full episode: clstk.com/4ibuX2Y

David Mokos from Expo has a few words for you ahead of Agent Conf 2026. He’ll talk about what changes when AI agents become the primary users of developer tools, from interfaces and documentation to evaluation and feedback. Find David’s talk and plan your two days in Warsaw: www.agent.sh/agenda

Meet Susmita Singh from Google at AgentConf 2026. She’ll show how to implement cryptographic memory signing, context-aware access controls, and adversarial memory sanitization to build agents that remember safely. Get a ticket: www.agent.sh/tickets

Come to Warsaw a day early for a free hands-on workshop with Amazon’s Giovanni Laquidara. Build a Software (re)Factory that works across Claude Code, Cursor, Codex, and GitHub Copilot. 📆 September 16, 2026 🕙 10:00–14:30 📍 Crowne Plaza Warsaw – The HUB Sign up 👉 www.agent.sh/workshop

Brownfield keeps the current app live but adds a boundary between two runtimes. Greenfield moves the risk to cutover. A third option starts greenfield, tests representative flows, and keeps a fallback into the native hosts. Download the guide: clstk.com/3TSBOEp

Anisha Malde is a React Native Developer Advocate at Amazon. At Agent Conf, she and Kourtney Meiss will use a football team to explain how agents coordinate, verify outputs, and catch failures. Tickets: www.agent.sh/tickets

Kewin Wereszczyński shares a view we probably all have at this point: nobody needs more vague AI talk. What we do need is a clearer way to judge where these tools are genuinely useful. Join us here on August 13 at 5 PM👇 youtube.com/live/eiGcqvj...

A QA agent should return evidence that a reviewer can inspect, not only a pass or fail. This clip from our last webinar covers five rules for keeping those runs reliable. For a full walk-through of the PR-to-device workflow and a live behavior-review demo, watch webinar recording: clstk.com/45cNynL

Kourtney Meiss works on developer experience across Alexa, apps, and games at Amazon. At Agent Conf, she and Anisha Malde will show how to coordinate agents, verify their work, and keep people involved. Tickets: www.agent.sh/tickets

People are tired of AI talk for a reason. Too much of it is vague and disconnected from real work. We’re sharpening our livestreams to focus on AI from a developer’s point of view: real workflows, where it saves time, where it adds risk, and where teams still need to slow down and think.👇