Kendra Connector

Connect Cohere to AWS Kendra.

Configuration

To use this connector, you must first configure AWS Kendra. Once you have Kendra configured, you will need to set four environment variables in order to use this connector:

,

AWS_ACCESS_KEY_ID

,

AWS_SECRET_ACCESS_KEY

and

KENDRA_INDEX_ID

.

The

environment variable corresponds to the AWS region that you have configured Kendra in. The

AWS_ACCESS_KEY_ID

and

AWS_SECRET_ACCESS_KEY

correspond to an IAM user configured with access to Kendra. The

KENDRA_INDEX_ID

value corresponds to the Kendra index that you would like this connector to search. The value can be found on the Kendra Index Settings page in AWS Management Console. You may optionally set the other AWS environment variables that are used by Boto3.

The environment variables can optionally be placed in a file called

. See

.env-template

for a full list of available options. This file can be copied to

.env

and modified. Options that are left empty will be ignored.

The Kendra index and returned data can be extensively customized in the AWS Management Console.

Finally, to protect this connector from abuse, the

environment variable must be set to a secure value that will be used for this connector's own bearer token authentication.

Development

Create a virtual environment and install dependencies with poetry. We recommend using in-project virtual environments:

To run the Flask server in development mode, please run:

The Flask API will be bound to :code:

.

Alternatively, load up the Swagger UI and try out the API from a browser: http://localhost:5000/ui/