HumanFirst & Cohere QuickStart Guide

COBUS GREYLING
November 2, 2022
.
3 MIN READ

To start using the HumanFirst [https://www.humanfirst.ai] / Cohere [https://www.cohere.ai] integration you will need a Cohere [https://cohere.ai] API key...in order to get your

To start using the HumanFirst / Cohere integration you will need a Cohere API key...in order to get your Cohere API Key.

The first step is to head to the Cohere website and click on get started.

You can register if you are a first-time user, or click on Log In should you already have a login.

Copy your API Key from the Cohere welcome page once logged in.

When going to production a production key should be created on Cohere.

The next step is to head to the HumanFirst Studio, you will need to create a new namespace which will point to the Cohere integration.

Click on the integration icon on the left panel, and secondly click on the Add integration button.

You will see the Cohere integration button, click on it.

Add a name for the integration, set the rate limit. If you use the  free plan there is a rate limit of 100 calls per minute (with 16 embeds per call) that means that it can only embed 1600 phrases per minute.

Copy the Cohere API Key into the bottom field and click on create.  Once you click on Create you should get a message that that integration has been completed successfully.

Click on the integration icon on the left, the integration detail is shown on the right, click on the Models tab.

Once you have clicked on models, click on Add model.

It is best to start with the small model...and click on Add model.

Now you can use the Cohere embeddings within the HumanFirst Studio for clustering and exploring data.

click on Create workspace.

I choose to create a black workspace.

I name my workspace Africa Facts.

Once your workspace is created, open the workspace.

Once the workspace is open, click on three dots and access the workspace settings.

Under Your Preferences, set the default latent space to the Cohere mode you have integrated to previously.

Now you can import data, and once the data is imported, it can be explored for semantical similarities according to varying granularity and cluster size.

Subscribe to HumanFirst Blog

Get the latest posts delivered right to your inbox