Openclaw + FREE GLM 5 API : Full Step By Step Guide

[0:00]Welcome back to another new exciting video. In this video, I will show you the full step-by-step process how you can integrate GLM5 free API endpoint with the Open Claw. So, in my last video, I told you that you can get this GLM5 free API through Modal Research, and they are providing this access for the free till April 30th, okay. So, this is the April 30th, you will get access, and uh it is it is completely free. And many of you asked me in my previous video that how we can use it through Open Claw. So, here is the full step-by-step video, follow with me, and it is a simple process that I will show you, okay. And over the YouTube, you will not find this kind of detailed video anywhere. So, please make sure to like this video and make sure to subscribe this channel also to get this type of content more. So, the first step is that to install the Open Claw. So, for that, if you are on Linux or Mac, then follow that uh Linux and Mac procedure. For that, um this is the command, install Open Claw. For Mac and Linux, this is the command. And for Windows, this is the command. And for Windows, there are two option, one is the Power Shell and another one is the CMD. I always preferred the CMD. So, that's why I will choose this CMD and copy this command and open your terminal and just run it. I have already run it, here you see, and I installed it, Open Claw. And after that, uh what you need to do, I have given this doc file link in the description. So, all of the comments, here you see that step one, step two, all of the step-by-step comments that is written here, you need to just copy it and run it, okay. So, we have already installed this first comment, uh this Open Claw. Now, the second step is that run the on board widget, okay. Now, this is the second command, just uh copy it and open the CMD, okay. Open the CMD and paste it. Now, make it large and press enter. Now, it will ask you the configuration. Here you see, I understand this is powerful and inherently risky. Continue. Yes. Continue. And on boarding mode, Quick start or manual. I will go for the Quick start. And uh config handling. So, here you see that I have already done the setup with the uh with the model. And this is the existing config. And uh now it is asking me, do you want to reset or use the existing value? So, I will go for the full reset. And if you are doing it for the first time, then you will not get these options, okay. But if you are using it for the second time, then you will get your existing config and it will ask you to the to do the full setup, okay, full reset. Now, I will select this full reset. Now, this is the model and auth provider. It is asking me to select. Now, I will go for the last option, okay. Select this up and down arrow key. Select this custom provider because I will use one Open AI compatible endpoint. So, select this custom provider. Now, it is asking me to uh put the base URL. This base URL you will get from the description. Otherwise, I have given this modal research GLM5 uh this page. And if you go to this page, you need to do the sign up and after that you will find this kind of base URL. So, make sure you copy this base URL up to V1, okay. Otherwise, you can copy it from the description. And just paste it here. This is the base URL. Paste it. Press enter. It is asking me to put the API key. So, for that, you need to come to this page. And here on the left hand side, you will find this create token. Click on this create token. Give a name here, like um testing, okay. Giving it name testing. Press enter. And this is my API key. So, here you see this is the API key. Copy this API key and just go to this terminal. Paste this API key. Press enter. Endpoint compatibility, it is the Open AI compatible. Press enter. Model ID. So, model ID is also present here. Just cross it. And here you see this is the model ID. Just copy this, okay. GLM5 AP8. This is a quanti version. And just paste it. Press enter. Now it will verify whether it is correct or not. And here you see that verification successful message I got. Endpoint ID and model direct. So, press enter. And model alias, you can give a name. So, let's say I am giving the name as the GLM5, okay. GLM5, and as this API endpoint is provided by Modal, so give the name as the Modal. By the way, guys, I am not promoting this Modal research anymore. I found this API endpoint free. So, that's why I am giving you uh this tutorial, okay. So, GLM5 model. Press enter. You can select any channel, Telegram, WhatsApp, but uh I am just skipping it for now. And skills status, yes. Configure skill now? No. And enable hooks? So, I'm skipping it for now, okay. And gateway service install failed. Gateway. Run PowerShell as administrator or rerun without installing the daemon. Okay. Let's wait for that. It will open a browser with the default port. Okay. Let's wait for that. Please open it, man. Please open it. Yes, it is opening. Yes, I got the error. It is a normal, uh that many people will get the error for the first time. Because we need to run another command, okay. So, what command we need to run? Open your CMD, okay. Open your CMD and just make it large. And the another command is this, this command you need to run. Open claw gateway hyphen hyphen port 18789, okay. So, when we will run this command, it will restart the gateway on this by default port, okay. All of the comments are given in description, you just copy it from description and uh paste it here and press enter. And now press enter. Okay, okay, okay. And here you see it is now ready, now let me just go to this browser and refresh it and you will see the magic, okay. Here you see in front of you, it is ready, okay. Now, it is ready to use. And if I just write hi and press enter, you will see that I will get the response back from this Open Claw, okay. So, let's wait for that. So, it is taking some time, that means there is some error maybe, so let me check that. Uh yes, we got the error. And model context window too small. So, let me show you it uh because this is a normal common error in case of Open Claw. Uh many time you can face this kind of error. Embedded agent failed before reply and model context window too small (4096 tokens). Minimum is 16000, okay. So, to fix this error, what you can do, uh go to your Open Claw setup. So, go to your C drive, or in case of Mac and Windows, you can follow uh the other paths. Actually, you need to find the .OpenClaw folder. And now in case of Windows, go to your user's profile. Go to your profile. And find the .OpenClaw folder. Where is that? Yes, .OpenClaw. Inside that, open this OpenClaw JSON file. So, open this OpenClaw JSON file. Open it, man. Yes, open it. And here you see. So, I have opened it. And here you see that context window is 4096 and maximum token 4096. So, that's why I am getting error. We need to change this context window to another uh another number, like 16000 or 32000, okay. So, I will do this 32000 for making it as a safe purpose, okay. So, this is the context window that you need to change. I will give this also in description, you can copy it uh from there. And here you see this is our GLM5. And this is the model endpoint. So, now save it, okay. And overwrite it. So, save this. And now again go to that terminal or command prompt. So, let me just uh refresh it. I can what I can do that I can again uh I I can again run this, okay. So, it will be better. Again Open Claw gateway port. Same command, okay. Now, this time, just go here and refresh it. So, now if I just write hi, press enter. Yes, here you see that I got the response. Hey, I just came online. Who am I? Who are you? Let me get my bearings here, checking what I, okay. So, I got this kind of response, right? So, in this way, you can use this GLM5 free API endpoint from Modal. And here you see these are the logs. Like today is February 15 and I got this recent logs. That API is getting hitted. And yes, okay. And if you found this detail step-by-step process helpful, please make sure to like this video, make sure to subscribe this channel to get this type of content more daily. And don't forget to like this video. Don't forget to subscribe this channel. See you guys in the next video. Thanks for watching. Bye-bye. Take care.