I loved every minute of it, but unfortunately, my time at Replay is over. The company decided to switch gears and will no longer work on test suite solutions. Hence, having someone as deep in testing as me on the team does not make sense. It is a huge bummer, but I understand

Hi everyone, here's an update on where we're at with our new direction: blog.replay.io/changelog-67:-… Over the next few weeks we'll be showing off demos for several cool and powerful new use cases for Replay. The first is full stack debugging, let us know what you think!

Is investigating and fixing test failures caused by your PRs annoying? Would you like a bot to do this for you? blog.replay.io/changelog-68:-…

Over the past couple of years, JS frameworks like Next.js and Remix have gone full stack. We believe it’s time for DevTools to join the party! blog.replay.io/changelog-69:-…

We're launching nut.new today to make AIs good at debugging and no-code development tools like bolt v0 Lovable Replit ⠕ better and easier to use. blog.replay.io/launching-nut.… veed.io/view/3663814d-…

We're excited to try the new 3.7 Sonnet more, but to fix bugs AIs need the right context. Here's a simple bug that stumps Bolt and 3.7 Sonnet. Nut has the runtime data to fix it with no trouble. bolt v0 Lovable Replit ⠕ Anthropic veed.io/view/f0c25c59-…

Fixing styling issues is a lot easier when you understand the constraints on an element's size and position. Here we use nut.new to fix a tiny search icon that v0 couldn't figure out. bolt Lovable Replit ⠕ veed.io/view/5634bf9e-…

Some bugs are subtle but will still totally break an app. Here we use nut.new to analyze and fix a bug in a vibe coded app that wasn't initializing its UI state right. bolt v0 Lovable Replit ⠕ veed.io/view/08963d22-…

Vibe coding can easily get React into bad states which are hard for the AI to fix. Here we use Nut to analyze and fix an infinite render loop caused by a missing useCallback. v0 bolt Lovable Replit ⠕ veed.io/view/44dae388-…

Today we're launching several new features for nut.new. You can build and deploy full stack apps, and Nut is now an agent. It QAs its work, writing tests and running and debugging them to get you a working app from the first prompt. blog.replay.io/nut-agent

nut.new now features a feed of randomly generated full stack apps you can use as a starting point for your project or just browse and try out, like you're walking through an arboretum. blog.replay.io/nut-arboretum

TODO apps are pretty simple and AIs are pretty good at building them, but they still mess up sometimes. Here nut.new (i.e. Claude) built a sort of broken TODO app, then identified and fixed its bugs: …4e46-b663-0b52f806433a.http.replay.io veed.io/view/e86b862c-…

Here's a voice journal app that Nut built with no human involvement. Voice transcription, entries database, search and insights all work great. v0 bolt Lovable Replit ⠕ app: …4d51-ada7-c16884bf18c5.http.replay.io fork: nut.new/app/cd21b1d5-2… veed.io/view/68d1ecef-…