Docs

AI

Let's take a look at how to use ai api call.

Deploying and integrating AI apps are time consuming and require a lot of expertise.🥵. But with BuouAI, you can integrate your AI models in no time ⚡ Let's see how you can deploy your custom AI models on GPU providers like Runpod or Replicate and integrate them within your project.

Deploying your own custom models.

You don't need to have your own custom models to get started. If your project doesn't need custom models, you can use the pre-built models provided by [Runpod] or [Replicate].

Essentially, all you need to get is custom endpoints to make requests to your models. Simply grab your model's endpoint url with one of the following services.

RunPod

Runpod Serverless.

Since we are all about shipping fast, we are going to cover how you can deploy your models on serverless platform so you don't need to worry about infrastructure and can focus on building your app. 🚀

In your Runpod dashboard, click on [Templates] > [+ New Template]. In here, you need to select the Dockerfile for your model. To learn more about how to create a Dockerfile, you can check out this Rundpod Worker Tutorial.
After creating your container template, head over to [Serverless] section, and click on [+ New Endpoint]. In here, select the template you created in the previous step.
Name your endpoint, select one of the available GPUs and configure how you want to scale up your instance. Then, click on [Deploy].
Your serverless endpoint will be ready in couple of minutes. After it's ready, click on your endpoint and get it's runsync endpoint inside the endpoint page. It should look something like this: https://api.runpod.ai/v2/12314874/runsync
In your BuouAI project, you can use this url to make requests to your model. Refer to the Runpod Features section to learn how to make requests to your model using RunPod.

Replicate

Get your API Key under your account settings.
After that you need to push your model to Replicate using Cog packaging library. You can easily do it by following the instructions on the official documentation.
After pushing your model to Replicate, you can use the API key to make requests to your model. Refer to the Replicate Features for more details.
[Optional] If you need more control on your deployed model, you can configure the model deployment. You can learn more about it from the official documentation.

Replicate Deployment.

We suggest keeping it simple and rolling with the default endpoints Replicate gives you with. 🚀

ChatGPT

You can easily integrate ChatGPT in your BuouAI project to create chatbots, conversational interfaces, and more. We handle all the boilerplate and retry logic for you👌.

Simply grab your OpenAI API key, pass in your prompt and get the response from the ChatGPT model. Refer to the ChatGPT Features to learn how to integrate ChatGPT in your BuouAI project.

Introduction Runpod