 process to move those files from point A to point B in the evening. So that is what I'm focusing on right now. I put together a first cut schema for a new wake word schema on our database that will facilitate all this stuff. Ken and I have been through that once. So I think what I have now is enough to go move forward with the upload piece of it. Ken, would you agree that it's at least at that point? Absolutely. Okay. So seeing that, and I shared it out to everybody in Matavos today, too, if you want to take a peek. Basically, I put a picture of the schema at the top of the document and some descriptive text underneath it. So let me know if you have any questions about how we're planning on doing that stuff. So I will start, since I have all that done, I'll start coding tomorrow. Get that API up and start writing it off to write the script that moves the data to the Lawrence server, to the Lawrence NAS, and that will be the first part of it. It will be that input part. And one thing that did come up as Ken and I were reviewing what exists today is that one of the things that didn't get moved to Salini from Precise, not from Precise, from Tartarus, is the ability to specify your own wake word and set the pocket-stink settings for that. We copied over those things from existing wake words, but you no longer have the ability to do that. You don't have the ability to change those settings on the UI. You have to change them on your device. So what I might do as part of this is resurrect that, just maybe a pop-up window on the UI that lets them edit that. Since we're working on the actual storage of wake words themselves and the inputs to them, so Michael, do you have any comment on that? No, not really. I'm just jumping in the middle here. It's a tiny bit of script creep, but it shouldn't be too bad. It's just something that, we have the ability, or at least had the ability for people to say, I want the my wake word to be my own. And there's actually what is there, Ken? Like, there's hundreds of them, I think, right now. Yeah, there might even be 300, yeah. Yeah, so it's definitely something people are using, but right now they can't update it. So I think I may just, I think the API is in place for it, but the UI screen there is not developed. So that might be something else I have a little input from you on Derek as well as, you know, what that would look like. Although I think it's just going to be a simple form, you know, that lets people input all the different parameters for the pocket things recorded. But so yeah, personally, personally, I think if it's not too much work, then it's a fun thing to add in. But I don't know that it's only that we have to provide, you know, I think people can use the devices. And the last week already people have a good experience with the default wake words anyway, because there's a lot of messing around with phonemes and thresholds and all that sort of stuff. So it's a good fun tweaking thing. But yeah, it's not too much work. And I know I'm sorry to everyone, anyone in the community is listening to this because I know it's not something that people want. But yeah, just in terms of what we getting to getting to the market production release and all that sort of stuff that's needed, you know. Yeah. And one thing that I really remembered from the last iteration of this is that there weren't any really good instructions as to when or why you would tweak any of the settings or what if they would help or not. So I mean, I'm sure if you somewhere in the pocket things documentation or something, but, you know, as far as, you know, basically as a screen with settings and you know, unless you knew what you were doing, I didn't be going to get anyway. But so I'll look into how the effort it'll be. It's going to be a big deal. But then maybe we can put it off. But just something I noticed as Ken and I were reviewing this that was another one of those little things that were omitted from Selene when we built it for the sake of time. So, okay, so I'm just looking at this here. The pocket things doesn't seem to use wake words in the same way that precise does. Is that true or false? True. Right. So that's a whole separate system, right? Like it doesn't really intersect with precise. I really know. Okay. It's available as a back as a backup. So if precise fails or can't run on the device for any reason, so they don't have the AVX instruction set on the TVU, then it falls back to pocket things. And so pocket things operate on phonemes. So it tries, it listens out for a collection of phonemes. When you have to define, use the CMU dictionary to define the phonemes that you want it to listen out for. So it serves two purposes right now. It's the backup and it also allows users to basically pick your own wake word. If they want to use, they want to use something that we don't carry. They could use, you know, whatever they want to use. So when you say users, do you mean like any end user? Or do you mean, you really mean like people who can download and modify the source code? You don't need to download. Pocket things are running on the device. So really what it is, you go in, what we used to do, be able to do is go into Celine and say, I want my wake word to be, hey, Michael. And you can, you can set the settings for that. We would, you know, put those settings on your device and pocket things would use that as your wake word. Apparently you have to edit the configuration file. At least not. Yeah. You can still do that if you SSH the device and change your convict file, but you can't do it via the Celine UI anymore. And you used to be able to do it with a torus UI. Yeah. Okay. Well, yeah, I think this is definitely something not a very good experience. It's not, it's not great. It's super old technology. It's not a great experience. It's good for a quick, you know, DIY kind of a thing, but, you know, DIY precise models are going to be significantly better. Okay. Well, then, you know, yeah, I don't think it belongs as part of this effort. Okay. Fair enough. Then I will, I will not worry about it. And I will just get the API up for the inputting data, the new way and storing it. And then we'll move on to the tagger. I think was the next thing we decided to look at. And then once we had those working, we'd move on to the new tool for manually submitting samples. So I think that was kind of the order that we had talked about as well. So we have a plan, which is good. And we have a lot of agreement on what we want to do. So Ken and I will work through a lot of that in a nice lengthy phone call on Friday. But yeah, so it's good. It's all good. And we'll get to her moving on it here tomorrow. Okay, great. So that's everybody's status, Michael. So if there's anything you wanted to interject before we stopped recording, then now would be the time. Nope, I'll have to go back and review the notes. See what I missed. But I can do that on my own time. Okay. Or you can listen to the recording of the meeting. Assuming it was successful. We'll see when I upload it. All right. So that is the meeting. And I will stop recording now.