Video: Office Hours: Cy.prompt() | Duration: 2576s | Summary: Office Hours: Cy.prompt() | Chapters: Welcome and Introduction (20.815s), Side Prompt Overview (197.51999s), AI Test Generation (363.89s), Enabling Psyprompt Command (537.145s), CyPrompt Usage Examples (618.06s), Reliable Cypress Prompts (830.475s), Optimizing Test Performance (1023.83496s), Self-Healing Test Optimization (1251.165s), SciPrompt Limitations and Features (1472.5249s), PsyPrompt Implementation Strategy (1614.0049s), Upcoming Event Notifications (1859.55s), Customizing Selector Preferences (1888.62s), Custom Login Commands (1994.61s), Addressing Final Questions (2089.035s), Element Selector API (2402.365s), Conclusion and Farewell (2508.9s)
Transcript for "Office Hours: Cy.prompt()":
Hi, folks. Welcome to this edition of office hours. We're gonna give just a little bit of time for folks to make their way into the meeting. So while we're waiting, give us a hello and chat. Tell us where you're joining from. I'll be here on stage, but I will be on mute for a moment, and, we'll come back on as soon as it's time to get started. Alright, folks. I think we have, given it two minutes here for folks to make their way into the meeting. So I think we're just gonna go ahead and kick things off. So, hello, everyone. My name is Jenna Beckett, and I am a senior DevRel engineer here at Cypress. And I'd like to thank you all for spending part of your Friday here with us at Cypress. This office hour session, as I said and you see on screen here, is all about putting Psyprompt to work. I'll show you how to go from test idea to runnable Cypress test in minutes. Oh, too quick on the clicking there. Now before we dive in, let's took, take a high level look at what we will be covering today. I'm gonna start off by making sure we are all grounded in common language about SIPrompt and talk a little bit about why we built it. We're gonna review the steps to getting started with the SIPrompt command, and then we will talk about how SIPrompt can be integrated into real projects by talking about a couple of common workflows we've seen it used for. We're gonna end on how to design your prompts for the best results and how to debug once you have SciPrompt in place. Finally, the most important part of this session is that we have left plenty of time at the end so that we can go through and answer any questions that you might have. Those questions, if you have them as they, come up during the presentation, please drop them over in the q and a panel. Once we get through the presentation piece, then I'll pull up that q and a panel, and we'll go through them and answer them each live. So, we're gonna get started by defining a common understanding of the side prompt command. Side prompt lets you describe test steps the same way you would explain a flow to a teammate. Cypress interprets each step, chooses selectors, and generates commands, and then runs them, which all sounds great until something goes wrong. In many AI testing solutions, you lose visibility into the code that is actually being ran. With Cypress, you maintain full visibility into that code, and I'll talk about exactly where you can see that, throughout the presentation. You can visualize specifically within the local app when you are designing the test and you're looking at the command log, you'll see the prompt steps written in, plain natural language as you used, and you can expand it to see the actual Cypress code that was generated. Or if you are running the command in a, like, headless test run-in a CICD pipeline, and you are recording, then you can actually see those same command steps and definitions in, the command log of a test replay. Finally, once you have your test, running, we designed this as a choose your own adventure solution. So you can leave prompt, in place and elevate it through version control, or you can export the code for maximum control. And we'll talk more about that. Now you may be asking, why did we build this? Why do we build it now? Well, our answer to that is that our industry is evolving at a pace that rivals what we've seen in the early two thousands. We are all running on tight timelines and even tighter budgets. So maintaining a high velocity is more important than ever if you wanna keep your applications stable and your customers excited about your product. In speaking with hundreds of different testing teams, time was always a common thread when discussing bottlenecks, especially test authoring time. Our focus at Cypress has always been on making testing easier, so Cypress seemed a great way to help folks close the distance between the test coverage they need and the code that actually enables it. Now, of course, there are multiple ways that we could have achieved, expediting test authoring, but natural language gives us the benefit of being able to include other folks like product managers, for example, in the test authoring process that may not necessarily know how to write code. Now you may be saying, hey. There are already AI tools out there that can help with test authoring. And that's right. And here at Cypress, we basically use the gamut of them. In doing so and in talking to teams who have brought those kind of tools into their own workflow, we found a couple common pain points that we just couldn't get over. First of all, it was another tool and another syntax to learn. And frankly, in large organizations especially, that tool approval can be difficult and budgets even harder to secure. So not to mention that you still have the time you, have to take to learn the tool and get, you know, caught up to speed to be proficient. So we wanted to have SIPrompt be as frictionless as possible. Another pinpoint that we encountered is that while those external tools are great, versioning is something that was not built in, and you couldn't leave the prompts in place as code elevated into CICD pipelines. So our other big objective with this command was providing you a way to use AI that continues to provide ROI and only needs manual intervention when you deem it necessary. That's why we integrated this functionality as a command into the Cypress API so it can live with your existing specs and be versioned with the rest of your test code without any extra steps, and then, of course, help you achieve resilience in the face of fast moving progress. I mentioned earlier that this was a choose your own adventure kind of solution. So let's dive deeper into what that means. So talking about what we have here on the left as workflow one, this is when you use SciPrompt to generate tests, then export and commit the code. It is predictable and allows you to maintain full control over the test commands that are being ran in your pipelines while still getting that boost in the test authoring part of your development process. Now if we switch over to looking at, what's on the right hand side, this would be workflow two, and this is where you keep the Cypress command in place and committed. Cypress actually hashes the generated steps and, regenerates selectors when the DOM changes. So this is what is going to help you reduce flake and, you know, the amount of regular test maintenance that you have to do. And the best part of all of that is that you don't have to make an ultimate choice between one of these workflows either. You can mix and match both of these approaches in the same repo, even in the same test. Getting started with the Cypress command is pretty quick. There's only a couple steps, so let's walk through these. First, you have to enable the config flag that's in your cypress config dot JS or TS file. Then you have to authenticate with the Cypress Cloud. So authentication can look, like, a couple different ways. So first, if you're running in the local app, you could just log in through that local app, and that will allow you to have that, securely managed connection to the LLM. Or if you are running in CICD pipelines or a headless mode, you can pass in a valid record key. And that valid record key, is going to along, of course, with the record flag, it's gonna give you the same, level of authentication. And then after that, really, the next step is just to write your test your first test with the psyprompt command. There's no extra SDKs or plugins that you have to bring in. You work right inside Cypress API and the app that you're already used to. It will probably take you less than a minute to do all of this, assuming that you're already on the latest version of Cypress, of course, because that is one other piece to the puzzle. And then, of course, as long as you remember your password. That's my big problem. Adds a couple minutes each time. Alright. Now I've told you about the command. I've told you about why we built it. Now I wanna show some examples. So let's look at a simple login flow or a hello, Cypress, if you will. Each string in this is one action, so the steps are atomic. Cypress interprets the intent, selects targets, and runs the equivalent Cypress commands. The command log will show you both the prompt step in natural language and the generated code so you can inspect exactly what happened while the test was executing. You also maintain that full time travel debug ability as well. Now SIPrompt, and I think I've mentioned this already, but it is not an all or nothing solution. I keep saying that because that really, I think, is my favorite part of this command is that it is something that I can use strategically. So since it is a part of the same Cypress API you are already using, you can mix and match the cypromp command with other Cypress code. Here is a couple of workflows that we found are especially beneficial when you are just getting started with side prompt. So the first of which is when you are working with, you know, new, features or, like, a completely new test suite, starting from scratch, Psyprompt is really gonna be useful in that case, because it is going to help you get that boost in getting your initial coverage up and running. The next situation where this would be really beneficial, this being the side prompt command, is in areas where you are drowning in flake. Prompts will help stabilize intent while the UI is shifting. Happy paths and, like, for PR validations, these are also great solutions. If you're in working in an existing suite, then I would say as your happy path changes, it's a really good time to bring in side prompt. If you do have the investment time to go back and do a little bit of redesigning on your existing happy path, then it can just give you that additional stability, in those workflows, especially knowing how essential they are to keeping your development process moving forward. For folks who are fans of DDD, you can write given when then syntax, and you can put that right into the prompt command without having to maintain separate step definition files. And then, finally here, when you are using, you know, tests that cover, like, repetitive steps, So, one thing that pops to mind for me is, you know, testing across a web viewport size and a mobile viewport size. Then you can reuse the same prompt and just pass in dynamic values. So now that we've talked about, you know, the prompt command itself, we've talked about getting started, one thing that I mentioned here is that making sure your prompts are clear and provide enough context is gonna be essential to, the benefit that you receive from this command. So, this is all, like, a big way of saying prompt clarity drives reliability. And here's a couple ways to make sure that reliability is achieved. So first, you wanna keep each step atomic, so one action per step. When you are trying to reference exact values like text, or maybe some type of unique selector, then, using quotes to make sure, Cypress knows that those are exact values we should match against will definitely be helpful. Give context when there are multiple similar elements. So, like, if you're saying to a coworker, go get that input in the form. If your coworker's response would be, what input? Well, the LLM is gonna have the same thought. So making sure that you're providing enough context to locate the specific thing that you're targeting. And and one way that, you know, I didn't cover this in the the first item, but I would like to revisit is you can keep that clarity if you avoid using ambiguous pronouns like it. Make sure you're being very specific, instead of, using those kind of ambiguous terms. I would also say that it is important that you understand that you must use absolute URLs for your visits. And the only exception to that rule is if you have a base URL, which I hope you do, already defined in your Cypress config file, then you can absolutely use, relative URLs, and it will build off that base defined in the config. And last thing here on screen, I'm gonna go back to context because this is such an important part of, designing your prompt steps. In fact, recently at Cypress twenty twenty five, a colleague of mine, Eric, did a Psyprompt workshop, and he mentioned a really good framework for looking at each command step. So his framework is that he's always gonna list the action first, the target second, and the location third. So something like, focus in on the email input in the login form. That would be a really good example of following that flow. If you follow that pattern for each one of your command steps, you should be definitely providing enough context to get that reliability that we're talking about. Another good tip for using side prompts, and I've mentioned this a few times, but let's dive in a little deeper here, are these placeholders. So placeholders actually solve two unique problems when we're dealing with side prompts. The first is privacy, and the second is performance. So let's tackle the first one, privacy. When you are using, placeholders, you can use them to keep sensitive information from being sent over to the LLM. Resolved values for placeholders will not be passed to the LLM, so this is really important, especially if you're passing in, like, login credentials, like a password. Another thing about these placeholders is they help you achieve performance by letting that cache stay, established and not have to be regenerated unnecessarily. So let me dive into a little more about what I'm talking about there. So let's say that you are, executing a common set of steps on a group of data. Maybe you're pulling it in from a fixture file. And you might have a couple, you know, dynamic pieces of information. If we're looking at our our example on screen here, I will just use, username and password as my example. So I'm gonna pass in a different user, a different username, a different password every time I run this prompt command as long as the actual value of the placeholder itself. So in my example that you're looking at on screen, placeholder called password, that has a value of super secret pass. If password stays the same, not pass, not cred, not any other value I could use for this placeholder, then the test steps stay the same. The cache stays established and is leveraged for the run to achieve as performant a test execution as possible. When test steps actually change, though so let's say I did have a placeholder value that changed from pass to password in, different instances. When that change happens, we're gonna recognize that a difference in your test steps have been established, and we need to go regenerate that cache. So to be able to get your test designed here probably in the local app, get the cache established, and then realize those performance benefits out in the CICD pipelines, then you need to make sure those placeholder values are staying consistent and using them whenever you need to pass in dynamic data. Now, when something you are working on, and you're using side prompt and an output maybe surprises you, the first thing that you're gonna do is go to that command log and check to see the exact selector and command that was generated. Again, if you're running in CICD pipelines, you will see that information in your test replays represented the same way you would see it in the local app as well. I would also say in addition to, your CICD runs that we've recently enhanced the experience with side prompts so that you will see a self healed tag. So, if we have, established a cache, we have tried to run your test using that cache, and for some reason, the UI shifted, maybe a selector change. We need to go back, regenerate the cache within, the CICD run, and try the test again. We're gonna label that for you as self healed. So in your test replays, you can actually visualize, hey. Something changed here. I have insight that these generated steps were adjusted, and it might mean something that you just wanna look into. Now the video that you just saw actually want to play that again because what happens when you're designing the test, right, and the steps that you provide do not give enough context? This video actually shows you the workflow that happens in that situation. We generate a dialogue for you, and I'll play this one more time so everybody can see it, where you look at the step that is not, specific enough. You can provide it more details, save that, and what's gonna happen is it actually saves those changes back to your spec file in your IDE, and it continues on with the run. And that's the most important part. Right? You're able to give it more context, but not lose all the progress you've made. Well, assuming you're running hundreds of tests locally, you don't wanna have to sit through and wait all that time again. This is gonna allow you to do that. Now I did get into self healing just a little bit in, our last slide, but let's let's dive in deeper. So I wanna make sure everyone knows that the regeneration is scoped. So Cypress does not recall the model on every step of every run. What happens is you develop your test locally. Once it is running and it is green, we establish a cache. That cache is built off, the name of the test, the spec file that it's in, the definition of the steps that are in the sciprompt command. We will only go back and regenerate the steps if we try to run the test and, it actually fails. Like, the selectors have changed, something about, you know, what we were targeting previously is no longer working. The other situation where we would go back and regenerate steps, of course, is if your cache is invalidated because the name of the test changes, the step files change, the location where it's at, so in a certain spec file changes. Those are all situations where we would also regenerate the cache. Being able to have this cache in place allows you to, again, run this at scale, seeing, no performance hits because we're hitting the LLM every time that cache is established. It allows us to go back, lean on that until it doesn't work anymore. And this is what really reduces those false failures, in your test runs due to just shifts in the UI that don't necessarily equate to your app being broken. I I've said this is, not an all or nothing solution, but I am going to continue to say that because, again, I love this feature of the command. So side prompts, again, can mix with regular Cypress code that can be included in, command chains, and they can exist in the same test in the same spec file with just regular old Cypress code. When you're thinking about when to, leave prompts in place versus when to maybe object to the code to achieve that full control, here's a couple pieces of advice. So I would keep prompts in place for navigation and interaction. And then, in certain situations, you might need to add targeted assertions in plain old Cypress code for clarity or because you have a very specific situation with specific values that you're trying to achieve. You might also mix it with regular Cypress code at this point, to cover the, parts of the Cypress API that are not exposed through the prompt command. When you're thinking about when to extract, again, I'll have a little bit more advice on that a little later. But I will tell you that, you know, I've had a lot of people mentioned to me, like, I want to be able to use side prompt, but I wanna keep this as dry and lean as possible. I don't wanna be writing even the same natural language steps over and over. Well, you can extract your prompt arrays, like your, natural language test steps, into something stored, in, like, a utility file. Or if you're keeping the same prompts in or steps in place for many of the tasks you're trying to achieve in a single spec file, you're you're working in plain old JavaScript. You could just define a, a function or, an array that contains those steps and then pass that into the prompt instead of a hard coded array of strings. I did mention that there's some limitations to sciprompt. I mean, I know everything I've said makes it sound pure magic, but let's ground ourselves in the reality of what the command can do today. So sciprompt does target end to end tests and runs in Chromium based browsers. It is optimized and tested for English language, natural language prompts, but the community has reported success with many different languages. I've heard folks trying it in Dutch, in Gaelic, and seen, success from using those in the natural language steps. At this stage, not every Cypress API is exposed. As I mentioned, we will be working hard, of course, over, the next, few months to make sure that we are expanding this feature, and getting it to a GA state for you all. But your feedback can certainly help us know what the community really wants to see prioritized. So there is a feedback button for Psyprompts in the local app. You will see it as you are using the command, and you'll see it right there, like, in the open view, as you're designing your test. For a full list of all of the APIs that are exposed to the sciprompt command today, go take a look at the sciprompt docs. There's really great information over there about that, and I have that linked in the docs tab of this, event. The last thing, and I've mentioned this before, but I wanna reiterate, you do have to be logged in to the CypressCloud or use a valid record key to be able to leverage Syprompt. However, this is available across all Cypress tiers. And if you go in and you don't have a CypressCloud account right now, we recently extended our enterprise free trial when you start an account to thirty days. So that'll give you plenty of time to try SciPrompt, but also maybe explore a couple more of those features that can help your team get to, its testing goals quicker. If you have trouble selecting, which account or type of account would be best for your team, I do have a button in the event that says chat with Cypress. If you click that, it'll record your email address and will have someone from the Cypress team follow-up with you and schedule a time to chat. Now earlier, we started with a hello, psyprompt, and now let's bring it back full circle to a more realized SaiPrompt command. This is what it looks like for a happy path ecommerce checkout flow, clear steps, quoted values, assertions tied to visible user outcomes. Again, from here, you have those multiple paths, leave it in place, realize self healing, or eject the code for that ultimate control and governance. Folks, I will say that this particular test took me five minutes to write. I guess at this point, what's really holding me back in test authoring time is my own typing speed, so I'll put a little work in on that this weekend to get to the next level. This brings us to the end of the presentation, so here's what I'd like for you to do next. Take a spin through the command docs. You'll find, again, a link to that in the docs tab. Enable that config flag in your project and log in to your Cypress Cloud account via the local app. Then try using the prompt. Go look at one of your existing smoke tests. Try finding a place where you could use that for maybe, especially locating elements in the UI that you're gonna, be interacting with. Once you have a working, Psyprompt included tests in place, then take a little time and think about how you are going to implement this. Again, because it's that, choose your own adventure and you don't have to completely rewrite your test suite, you can be strategic about this. What areas of your test suite need the most help? And Psyprompt could help you with that. Or what areas are working fine now and you wanna leave the way they are. Make a plan for that, and then you can start integrating it on a wider basis. I will say a good rule of thumb is to export code in areas where you do need that strict review and governance and then to keep prompts in place in areas with high UI churn like new feature development. However, before you do all of that, I would love to hear what questions you have. So if you have not already, please go over to the q and a tab, drop your questions there, and I will stop sharing my screen and start looking through those questions. So let me go ahead and do that, get the screen down. Alright. Q and a. I don't see any questions yet, so I'm gonna pause here for a moment giving folks time to ask any questions that they have. Let me make sure I have all the filtering right, so I'm seeing all of them. So q and a tab, that's where you can drop your questions. I'm also looking at chat folks, so if you want to just throw something in chat, you can put it over there as well. But I'll pause here for a moment before we wrap up to make sure there aren't any questions from folks in the audience. Okay. So I see, a question here from Artie. So I'm gonna share that. Can I ask for a psi wait or psi intercept in psi prompt? Weights are currently achievable in the psi prompt command. Intercepts are not yet though. That is on, our team's radar for sure to get in because we have heard feedback on that a lot. So, not right now for the intercept, but definitely yes for the wait. And I'll give time here for any other questions that are coming in. While we are waiting, though, for questions to come in, I would mention to you that I said we did a PsyPrompt workshop at Cypress twenty twenty five. If you were not able to join that workshop or the conference, next week, we're actually doing a Cypress twenty twenty five rewind where we'll be replaying all of those sessions across our YouTube channel. So if you wanna get notifications, especially about that Psy prompt workshop recording, go on over to YouTube and subscribe, and you'll get those premier notifications. It's happening all next week. So you'll be able to see every session, every talk, every workshop that happened throughout the CypressComp twenty twenty five event. And there were plenty that we're covering, not only SIPrompt, but other AI solutions as well. Alright, folks. I'm not seeing any other questions in chat. Oh, wait. I spoke too soon. Here we go. Let me go ahead and approve some of these. So let's, share. We are using IDs in the Dom. Which info does the prompt look for? That's a great question. So, this is something that's actually customizable, because we are utilizing the, selector API. So, you can go in and in your own Cypress projects, tell Cypress, of the things that, you know, is in our hierarchy, our typical hierarchy of looking for elements. So we start with, like, data dash side, data dash, ID, those kind of attributes, and then work our way down to more generic things like classes and IDs. But, if you wanna see the full list, I'll grab you a link, and get that in chat while we're, talking here. But, you can actually go in and you can, customize that hierarchy so that it chooses the selectors that are preferential to you. And this is actually the functionality that, used to power the, selector playground that was in the app before the new studio experience came. So, it's still there. You can still use it. So, definitely, as far as, IDs, you can use those for selectors, within Psyprompt, but you can also go in and customize it so it it matches what your team is using. Okay. And I will stop share on that one. Artie, let me go ahead. I see another question from you and share that. Can we save side prompt for you for reuse? I want to save that side prompt login method that was made. Oh, yeah. So, a couple ways that you could do this here. So, if you are wanting to have, like, a, a login flow that maybe is repeated throughout your entire, test suite and you wanna reuse it over and over again between multiple, tests but still maintain a single source of truth, what you could do is you could actually create a custom command. Maybe you already have one for the login flow and put the psyprompt command inside your defined login custom command. That way, you have a single source of truth for those steps throughout your whole test suite, and you're getting the resilience self healing ability, as well that comes with side prompts. So that's definitely one way you could do that. Another way, and I kind of alluded to this in the presentation, is that if you have a common set of steps that you're gonna do in many different cases, you can define them in, like, a utility file, and you can, retrieve them and pass them into your prompt command. So we're still working in plain vanilla JavaScript. So what you pass into, the command does need to resolve to an array of strings, but it doesn't have to be a hard coded array of strings. Okay. Well, I think that is all the questions that we have. So, I will, let people go ahead. Well, I spoke too soon again, folks. You guys are awesome at this. As soon as I say no more questions, you're like, here's one more. Let's share this one. Okay. All of the actions which Cypress can automate is covered using Cypress. If not, what are those actions that are missing? Yeah. So, again, on the docs tab, if you go over there, and look at the Cypress docs, there's full list of everything that can be done along with examples. Off the top of my head, Cypress, Cypress, those are some of the big ones that are not, that people have mentioned that are not yet covered by the, Cypress command, right now. Remember, this is in experimental phase, so there's gonna be a lot of enhancements as we move towards GA. So, use that feedback button if there's some other command that I haven't mentioned that is really important to you, that's not exposed yet, to let the team know that's something you'd like prioritized. Let me go ahead, and we got another question. Does Cypress actually accept dynamic values like this after resolving an alias within? I'm not sure if it'll work as expected. Okay. Let me read that one more time. So, your mention of dot then here, makes me think that you're talking about a chain of commands. So, like, Cypress being the the beginning and then chaining dot bin onto it and performing some kind of, like, more custom action with just regular old Cypress, commands. So in that case, what sciprompt yields over to the dot bin is the last element that was, targeted in your prompt command. So if I have a number of different steps and the last one says click the submit button, then what is being yielded over is the submit button to the, dot bin. So, the other piece of your question here is the dynamic values. It does really accept dynamic values, as long as you're using those placeholders. So if you're not leveraging the placeholders, then though everything in the step is utilized as far as generating, Cypress commands. So it's sent over to the LLM and then from that, generates the test steps. Like I said, the placeholders, the fully resolved value for those placeholders are never passed over to the LLM in the process of generating, test steps. Okay. I think it it looks like we have some, like, a follow-up on this. So let me put that out here. Sci get, dot then. Ah, ah, okay. This this makes more sense. So if you are you could definitely do this flow. If you're trying to pull the data from the UI. And, yes, I would anticipate that it works. I would also anticipate you needing to use placeholders for that because if you're pulling something like name, that is something that will change, from test to test. So to make sure that your test steps do not change and the cache is invalidated, you'll definitely want to leverage placeholder values. I think, also, if you have the ability to pull that same data from, like, a fixture file, then you might not need the, like, leading part into Syprompt, the side dot get dot then, all that good stuff. You could just directly, you know, use prompts to, target those values with placeholders. Or, generally, like, you could also just use side prompt. Right? Like, you can pass to side prompt, you know, go, get the, you know, the value inside of this element and type it into something. I said that as two steps in one sentence. Remember to keep your steps atomic, one step per, line in the command. But, yeah, I think that there's a couple ways you could attack this flow. This, I do anticipate to work as long as you use those placeholders. But I also think you might be able, especially if you have the option of pulling that data from an alternate source to, maybe optimize this a bit. Next question here, I'm going to approve and share. Can we guide Psyprompt to select the locator as per our choice? Locator type, yes. Specific locator, no. I when and when I say locator type, I'm talking about, like, the data dash size, an ID, a class, something like that. So going in and leveraging that, selector API that I mentioned, you using it as it is or customizing it to your own, your your own team's preferences. But, yeah, that you can use it to to specify the type, but not the specific selector itself. Now, I do owe a resource to folks in the audience, so let me grab that for you. And that was our, documentation on that selector API. So, let's see here. Let me grab this link, and I'm going to drop it in chat. So, take a deep dive into, that element selector API. That'll give you a lot of great information on how to use it, how to leverage it for customization. If you were used to using the, Selector Playground, then I also, would recommend checking out this part of our documentation, because it does give you, really useful information on, you know, how things worked with that old Selector Playground and what to do with this revamped version of the element selector API. Alright, folks. I think that is the end of our session here and the end of the questions that we have. So, I am gonna go ahead and, wrap things up. I wanna say a huge thank you to you all again for spending a part of your Friday here with us at Cypress. Hope you got some valuable information out of this session. And, if you are wanting to rewatch the session, we have dynamic replay turned on. So as soon as the session concludes, you click the event link, you'll be able to get access to the replay. Usually takes about thirty minutes for it to process before it's available. But, it's out there and ready for you to use. So, thank you again. I hope you give Psyprompt a try. If you have any additional questions on the command, please leverage that feedback button inside of the app, or you can always reach out to the Cypress team at support at cypress. Io. We appreciate it and hope you have a wonderful weekend. And, of course, happy testing, folks. Have a good one.