Creating ChatBots & Inference API's

This guide will help you get started with our innovative suite of technology solutions designed for the digital era.

In case you dont have an account with us, follow the steps here or follow the steps below to get started.

Navigating the Meda Dashboard

Once verified, you'll have access to your personal dashboard

Those are the differnt sections you could navigate

Our Interface is very straight forward, it currently allows you to add data sources;

Adding data sources

Click the Create Source Button

The reset here is straight forward;

We support sources that include Web page urls, PDF documents (For now) and also plain text that you can paste directly in our interface;

We provide AI powered means of improving the uploaded content, making it easier to generate better information;

The AI models availed to you depend on the plan you have subscribed to, the more advanced the plan, the better the tools availed;

Data Source Management

You can also edit and delete the data sources according

Creating Assets (ChatBots & Inference)

For Both Chatbots & Inference API',s its required the creation process is slightly the same however their configuration is different.

Using the data sources created you can now create Inference API's, ChatBots, Summarization Endpoints Etc ;

1. Click Create Asset, and choose what kind of asset you wish to create

Choose ChatBot In Case you want to create one

1. Fill in the necessary details of the asset as shown above next provide a system prompt (this is required for much of the AI powered services we provide)

1. Lastly select the necessary sources that the asset may need to reason, you can change them later whenever you can;

Enable those sources that you will need to use

On save you will be able to view your saved Assets at the Deployments Tab

For an API Endpoint

For a Chatbot

ChatBots: Configuring a Chatbot

After creating a chatbot, as above you cant use it immediately, more configuration is required for it to be usable; (You will notice that each Chatbot instance has an indicator);

🔴 - Means the chatbot has no configuration yet

🟢 - Means the chatbot can be exported (shared)

This requires configuration

On clicking configure, you can now access this interface;

1. Changing the Look by adjusting the colors

Change the interface lok

1. Adding help links (more resources to help visitors of the website)

You can add as many as you want

Adding more links

1. Updating particulars like (Organization Information, welcome message, display position etc)

Updating details

Adjusting positions

1. Sharing the ChatBot (Using it in a website)

Green means all is well

Copy and embed snippet

You can now paste the snippet anywhere in a website

Sample code snippet

Sample Conversation

API's: Exporting & Accessing the API's

Currently, to export an asset (Say an Inference API), you require the asset ID and your Account API details;

You cane access the ASSET ID from there;

Account API details;

Now you need to get the API endpoint URL which you can copy from the overview panel (On the deployments section) of your account ;

API url endpoint

The same API Endpoint is what you will use for all chat API,s you have created, its different from that when interacting with hosted Machine Learning models (see here)

Also When you click on the Quick Guide button, the following will be shown to you, this interface alsp provides you with the same API URL endpoint that is specific to your account (you will use it for all interactions);

It will require some skills with programming in any language so as to be able to use the API, or any other assets like ChatBots.

How to access the API's programatically

Andromeda Labs also allows you to access your assets via a dedicated endpoint, that consumes the necessary data the endpoint expects ; For more information on this ; visit here