{"id":3757,"date":"2023-06-25T11:13:29","date_gmt":"2023-06-25T11:13:29","guid":{"rendered":"https:\/\/picockpit.com\/raspberry-pi\/?p=3757"},"modified":"2023-07-03T10:33:23","modified_gmt":"2023-07-03T10:33:23","slug":"raspigpt-ai-assistant","status":"publish","type":"post","link":"https:\/\/picockpit.com\/raspberry-pi\/raspigpt-ai-assistant\/","title":{"rendered":"RaspiGPT – Build an AI assistant"},"content":{"rendered":"\n
\"RaspiGPT<\/figure>\n\n\n\n

With the meteoric rise of language models such as GPT-4 we wanted to bring the power of AI to PiCockpit. So our developer Chloe designed the new RaspiGPT Tool for PiCockpit<\/a>, which is a GPT based AI assistant.<\/p>\n\n\n\n

It\u2019s purpose is to answer Rasperry Pi related questions right here in PiCockpit<\/a>. Of course if you need to know something else, it will gladly answer your question but with a quick reminder that it is meant for Raspberry Pi and technology related topics.<\/p>\n\n\n\n

We were amazed how easy it is to set up such an AI assistant so we want to show you how to create a similar application yourself!<\/p>\n\n\n\n\n\n

1 How does the AI assistant work?<\/h2>\n\n\n\n

The app is quite basic but it contains just about all you need. There is a text field to enter prompts and buttons to rename, report, delete or to create chats.<\/p>\n\n\n\n

\"RaspiGPT<\/figure>\n\n\n\n
\n

Note<\/mark><\/strong> You need an OpenAI API Key to build your own version of this app.<\/p>\n<\/blockquote>\n\n\n\n

2 Install dependencies<\/h2>\n\n\n\n

We are going to be running the demo on a Raspberry Pi 4 but any other system that runs node.js will work exactly the same. Installing the dependencies for our AI assistant is pretty straight forward.<\/p>\n\n\n\n

First we need to clone the Repository from GitHub.<\/p>\n\n\n\n

<\/circle><\/circle><\/circle><\/g><\/svg><\/span><\/path><\/path><\/svg><\/span>
git<\/span> <\/span>clone<\/span> https<\/span>:<\/span>\/\/github.com\/pi3g\/GPT-AI-Assistant.git<\/span><\/span><\/code><\/pre><\/div>\n\n\n\n
\"node.js<\/figure>\n\n\n\n
Since I am running the example on a Raspberry Pi 4 I need to download the ARMv8<\/em><\/strong> version.<\/p>\n\n\n\n
To extract and install node.js run<\/p>\n\n\n\n
<\/circle><\/circle><\/circle><\/g><\/svg><\/span><\/path><\/path><\/svg><\/span>
tar<\/span> <\/span>-<\/span>xavf<\/span> <\/span>node<\/span>-<\/span>v18<\/span>.<\/span>16.1<\/span>-<\/span>linux<\/span>-<\/span>arm64<\/span>.<\/span>tar<\/span>.<\/span>xz<\/span><\/span>\ncd<\/span> <\/span>node<\/span>-<\/span>v18<\/span>.<\/span>16.1<\/span>-<\/span>linux<\/span>-<\/span>arm64<\/span>\/<\/span><\/span>\nsudo<\/span> <\/span>cp<\/span> <\/span>-<\/span>R<\/span> <\/span>*<\/span> <\/span>\/<\/span>usr<\/span>\/<\/span>local<\/span>\/<\/span><\/span><\/code><\/pre><\/div>\n\n\n\n
Of course the filename might differ for you depending on which version you downloaded. <\/p>\n\n\n\n
Use these commands to check if node.js and npm were installed correctly.<\/p>\n\n\n\n
<\/circle><\/circle><\/circle><\/g><\/svg><\/span><\/path><\/path><\/svg><\/span>
node<\/span> <\/span>-<\/span>v<\/span><\/span>\nnpm<\/span> <\/span>-<\/span>v<\/span><\/span><\/code><\/pre><\/div>\n\n\n\n
These should output the correct version or an error if something went wrong during the installation.<\/p>\n\n\n\n
\n
Note<\/mark><\/strong> Further instructions can also be found in the README file<\/a>.<\/p>\n<\/blockquote>\n\n\n\n
Now we can install yarn using the following command.<\/p>\n\n\n\n
<\/circle><\/circle><\/circle><\/g><\/svg><\/span><\/path><\/path><\/svg><\/span>
npm<\/span> <\/span>install<\/span> <\/span>-<\/span>g<\/span> <\/span>yarn<\/span><\/span><\/code><\/pre><\/div>\n\n\n\n
With this, all the dependencies are installed and we can proceed with building the app.<\/p>\n\n\n\n
3 Build the app<\/h2>\n\n\n\n
Let’s see how to build the AI assistant.<\/p>\n\n\n\n
Navigate to the API Keys page of your OpenAI account and create a new API Key.<\/p>\n\n\n\n
\"create<\/figure>\n\n\n\n
You can name it something like “RaspiGPT-Key”<\/em>. Make sure to copy the API Key to your clipboard.<\/p>\n\n\n\n
This needs to be pasted into line for of the “nuxt.config.ts”<\/em> file.<\/p>\n\n\n\n
\"AI<\/figure>\n\n\n\n
4 Run the AI assistant<\/h2>\n\n\n\n
At this point all thats left to do is to build and run the AI assistant. Hence we need to run the following commands (make sure you are inside the “gpt-demo”<\/em> directory).<\/p>\n\n\n\n
<\/circle><\/circle><\/circle><\/g><\/svg><\/span><\/path><\/path><\/svg><\/span>
yarn<\/span> <\/span>install<\/span><\/span>\nyarn<\/span> <\/span>build<\/span><\/span>\nyarn<\/span> <\/span>preview<\/span><\/span><\/code><\/pre><\/div>\n\n\n\n
The last command will host the AI assistant locally so to access the App open “http:\/\/localhost:3000”<\/em> in your browser. A cool feature is that the answer is streamed as it is being generated, so you can read along.<\/p>\n\n\n\n
Watch this video (coming soon) to see the AI assistant in action.<\/p>\n
Now we need to install node.js, npm and yarn. Go to the node.js downlad page<\/a> and download the required version for your system.<\/p>\n\n\n\n