n8n 2.0 Just Dropped! (Everything You Need to Know)

Christmas came early. It's finally here. NAM V2 has officially dropped and it's an absolute beast of an upgrade. New features, major bug fixes, performance boosts, security upgrades, and yes, that long standing sub workflow bug, it is gone. If you've been building in NAT M for a while, this is the kind of update that makes you fall in love with automation all over again. But first, let's talk about the interface. Because the moment that you open Anatom V2, it just hits a little bit differently. Smoother, cleaner, more focused. They've made some relatively bold layout decisions without groundbreaking and changing everything. So everything's broken. I am all here for what they've done. Firstly, everybody just thinks that V2 is a cosmetic update. It's a fantastic cosmetic update, but they are wrong. The UI changes, they're great. The save speed, amazing, drastically better. But here's one thing that actually changes how workflows behave. I'm going to show you it at the end after we've gone through the sexy stuff. I know that some of you will have built a workaround for a sub workflow book before you already know exactly what I'm talking about, if you have, and I'll get to that shortly. But firstly, let me show you what immediately is different when you open v2. It just feels different. First, the nodes look different. They're flatter, they are smoother, they are more futuristic. No more white or gray outlines. Everything feels cleaner and more intentional. And honestly, this matters more than what you think it would when you're staring at a canvas with 50 nodes on it. Visual clarity isn't cosmetic, it's actually cognitive load. Faster parsing, easier debugging, less mental friction. The second thing that I would say to you, just to show you that quickly, this is what it used to look like now hurts my eyes. Like I don't want to look at it. Like I really don't. So I'm going to go back to V2. But when you look at this, okay, when any node that can take a few seconds to launch, right, Would not give you any execution feedback, you don't, you'd be like checking the executions tab. Watch what happens here. If I say hello to this agent, I've put a wait note in there so you can see this. Look at that weight node, okay? Anything that takes a few seconds to execute this circular or square animation in progress is genius. And here's why it's genius. Because it solves a psychological problem, not a technical one. In v1, when a node took 10 seconds you'd start wondering, is it stuck? Is it working? Did I do X correctly? Should I cancel it? Should I go and check something? In V2, you know it's working. The animation replaces anxiety with certainty. Be honest. How many times have you stared at a workflow wondering if it froze or if it's just slow? That moment where you're hovering over the cancel button, but you're not sure yet if you've lived that. You already understand why this animation matters. But here's where things get interesting. Saving is now instantaneous. This isn't full auto save yet. That is coming hopefully on December, the December 15th. But I'm hearing that it might be January anyway. Regardless, saving in V2 feels completely different. In V1, you'd click Save Micropause Confirmation. In V2, you click Save. It's instantaneous. Already done. There's no interruption of flow anymore. And here's why this matters more than it sounds. UIT saves dozens of times a session. You should be hitting Save after every edit that you make. That micropause. It could potentially break your thought process. V2 removes that friction entirely. It sort of removes. It doesn't feel like typing or submitting a form like it did in V1. It's literally just boom, and it's done. However, the key isn't that this is how you know Autosave is ready. Autosave only works if saves are invisible. V2 just proved the foundation exists, and I'll show you right now. This is V2. You move it, you hit save. Instantly done. This is V1. You move it, you save. Takes time. That's how you know that they're ready for autosave. Publishing is now also versioned. In v1, you add a simple toggle, activate, or deactivate. In v2, publishing is treated like a versioned action. You hit publish, name the version, add the notes and the initial draft, or public testing, etc. Whatever you want to call it, and that version goes live. If you want to unpublish, you don't hit Publish again. You open the workflow menu, go to Settings and select Unpublish. So you go in here, hit unpublish on the screen there. That's the difference. Whereas in the old one, it was just a toggle, literally turning on and off, turning on and off of the. Of the flow. So that's another slight change. Different navigation where you find it. It's more intentional, more deliberate. Right now, there doesn't seem to be a way to attach different versions to the same public URL that might evolve later, but for anyone deploying workflows publicly, especially teams and teams of groups of people, this addresses the problem of accidental production changes with no context. Another feature which is incredibly basic but powerful, the Settings panel. It's now available in the sidebar. In V1, you'd have to go in, you'd have to click your icon, you'd have to go down here, click this and hit Settings and then you will load up the tab. In V2, we don't have to do any of that. We literally just hit settings and they're all of the opportunities and all of the options pop up. Okay, you literally just click it. Once you're inside, everything looks familiar. Same options, same structure. And that's actually a smart move. Nan avoided the risk of breaking muscle memory while modernizing the visuals. But here's the part that matters most now. One final thing before I show you all how you can upgrade safely, is the breaking change that actually affects how workflows behave. This is the one that everyone needs to understand. The bug has been around for years. In V1, when you used execute sub workflow, the data coming out of that node was just the same data you sent in if it was over 60 seconds. If there was a wait node somewhere or something that took longer than 60 seconds, the data that it got sent in the primary workflow was exactly the same. The sub workflow's internal result effectively ignored. Which meant no true human in the loop logic, no waiting for approval inside a workflow, no meaningful data returned after long ways. You knew the sub workflow changed the data, but it pretended like it didn't. It was like the workflow was gaslighting lighting you. And for years people built workarounds, external approval systems, manual polling state stored via databases or global variables, artificially flattened workflows just to avoid execute sub workflow, all to compensate for one missing return value. What actually happens in V2? I tested this myself and here's the setup. This is the parent workflow. Okay, we'll load this up. This is the sub workflow here. Drag that out for you. My sub workflow one. And inside of this sub workflow I've got a 65 second wait because 60 seconds was the old one. What we're going to do is here in the set fields. Well, actually, let's do it the other way around. Here in the, in the set field section I've got value. Please subscribe. Please subscribe if you enjoyed the video. And then in the sub workflow, the value is thanks for Subscribing. Okay, so I'm going to execute this workflow now. Okay, I'm going to hit it. Executing. You can see the beautiful execution of sub workflow there. Running. You can go into the sub workflow, right, and look at executions because this is obviously how they work. And you will also see that this is still running inside of here. And once those 65 seconds are up, I'll show you the result. And as you can see on the screen now, the execute workflow is there. The value that entered was Please subscribe. You can see it here, the value that was returned. Thanks for subscribing. Thank you Nan for fixing this sub workflow issue. Because it has been a nightmare since day one. Okay, so now it waited for the the submission of the sub workflow. You know, in V1 the parent would receive the please subscribe or whatever it was. In V2 it now receives the actual direct response which was thanks for subscribing the sub workflows result is actually return. And when I saw that output there was this moment of disbelief. Wait, it actually came to life, like they actually did it. Because this exact limitation has existed for years. But then the disbelief turned into excitement because suddenly workflows that were impossible before or just extremely tricky are now incredibly easy and possible. What this unlocks. You can now run a human in the loop inside workflows, wait for approval or rejection. Return that decision branch logic into your main workflow afterwards. Clean approval flows, long running decision processes, modular workflow designs that actually behave predictably. All the things that people assumed execute sub workflows already did. It finally does them. So just quick guys, hit me in the comments. What sub workflow workarounds are you finally deleting? External state management, Manual polling? Whatever it has been that has been plaguing you for years, let me know, because here's what this means for the future of NAN. After using NAN 2.0, here's the honest takeaway. The UI is much cleaner. It's exactly what I wanted. More modern. Visual clarity matters when you're building flows like this. Execution feedback matters. Safe speed matters. These aren't cosmetic. These are cognitive load reductions. The platform still feels familiar. NAN didn't force you to relearn everything. They protected your muscle memory while modernizing in experience. That is a smart move. The sub workflow fix is a massive functional win. They should have done it ages ago, but they've finally done it. This isn't hype. This is an engine level behavior changing. If you've ever struggled with sub workflows before this alone justifies upgrading. But here's what this really tells us. V2 isn't a redesign for the sake of redesign. This is NAN paying down technical debt and stabilizing its foundation. What does that sell better autosave, more reliable long running workflows, Enterprise grade execution confidence. This is NAN growing up if you like and if you're building serious automation, V2 feels like the version of Nan that's been building towards for a while. So now the most important thing when it comes down to updating to NAN v2 to is please do not do it right away. Let the bugs get fixed first. That was in my previous video. But seriously, don't when it comes down to this, what you need to do right now is you need to go and have a look at the migration report and this migration report this I'm back in V1 because this is where obviously you'll be. You go to settings, you hit migration report and inside of here you will have a compatibility report for version 2.0 and this report. Inside of this particular instance I've got many including self hosted. I've got 144 of 145 of my workflows are already compatible with version 2.0 so I have not much to worry about inside of this one. There is a few other ones in in some of the other ones but if I go into this tells you what it is, right? The pyodide based Python implementation in the code has been removed and replaced with a native Python task runner implementation. So I click the workflow, it shows me it's a deactivated workflow. Apply K means clustering algorithm. This is the node that's affected. If I click that, it'll take me straight into the individual node that's there and it'll show me exactly where the problem is. In this particular case for this individual particular problem, all I've got to do is change it to the native one, come out of there, save it. Once I've saved it, if I go back to this personal settings and refresh, go to migration report, no more workflow issues. Now there's a low, high, low, medium and high level rating that you get from some of these things. Some of these won't really matter, right? Some of them won't. I don't want to encourage you to like cut corners really, but you know they're already compatible with 2.0. Some of them will have low impact issues, but when they do, some of them you'll need to fix others of them you know, they're just basically like not optimal in, in Nan's eyes, which, you know, I can live with that. But I've just shown you how you find your workflow issues. You might come in here and you might see 50 workflows inside of here. Do not be scared. Go into each individual one and it will take you directly to the node that is affected. And then you will have to find the solution. Some a lot more technical than the one I've just done. But I'm giving you, I'm showcasing to you how you fit, how you update. I personally would not update. You can update your workflows, but I would not update to version 2.0 until the release. The first patches 100%. So here's the bottom line. NAN 2.0 is out. And it's not just a UI refresh. Although it's amazing, the nodes look cleaner, the execution feedback is better. Saving is almost instantaneous. Ready for autosave publishing is kind of Version settings are easier to access. The platform still feels familiar. The sub workflow fix has been resolved. What more did you guys want out of this? Let me know in the comments, like subscribe. My name is Michael Elliott. Thanks for watching.