Wickr Bedrock Bot

A Wickr bot integration with Amazon Bedrock that enables interaction with Bedrock's foundation models through Wickr. This sample demonstrates how to use TypeScript for Wickr bot development and deploy serverlessly with AWS CDK.

Project Structure

.
├── bin/                   # CDK app entry point
├── bot/                   # Bot implementation
│   ├── src/               # Bot source code
│   ├── scripts/           # Bot runtime scripts
│   ├── types/             # Type definitions for Wickr bot module
│   └── Dockerfile         # Bot container definition
├── lib/                   # CDK infrastructure code
└── test/                  # CDK tests

Getting Started

Prerequisites

Node.js 20 or later
AWS CDK CLI
Docker
AWS Account with appropriate permissions

Additionally, the host machine you're deploying from needs at least 2 CPU cores and 4GB of RAM to successfully compile the code and build the containers.

Configuration

Copy the example configuration file:

cp config.example.yaml config.yaml

Update the configuration in config.yaml:

# The AWS account ID where resources will be deployed
account: "xxxxxxxxxxxx"
# The AWS region where resources will be deployed
region: "us-east-1"
# Flag to indicate if this is a development environment
# This will disable VPC flow logs, allow remote access to ECS containers for debugging, and 
# set the retention policy on infrastructure resources to DESTROY for easy clean up
isDevelopmentEnv: true
# (optional) ID of an existing VPC to use for ECS Fargate
# delete this if you want a new VPC to be created automatically
vpcId: 'xxxxxxxxxxxxxxxxxxxxx'
# The ARN of the AWS Secrets Manager Secret containing the bot username and password
credentialsArn: 'xxxxxxxxxxxxxxxxxxxxx'

AWS Wickr Setup

Before deploying, you must set up AWS Wickr and create the necessary users. Follow these steps:

Create a Wickr Network (skip if using existing network):
- Open the AWS Management Console and navigate to AWS Wickr
- Choose "Create a network"
- Enter a network name and select your plan
- Complete the network creation process
Create a Regular User:
- In the AWS Wickr Networks page, select "Admin" for your network
- Navigate to User > Team Directory
- Choose "New User"
- Enter the required information (email address is mandatory)
- Create the user and note that they will receive an invitation email
Create a Bot User:
- Navigate to User > Bot Management
- Choose "Create New Bot"
- Enter the required information
  - NOTE: Username must be lowercase, cannot contain any spaces or special characters other than -, and must end with "bot" (e.g. bedrock-bot)
- Make note of the username and password - you'll need these for the next step
Store Bot Credentials:
- Navigate to AWS Secrets Manager in the console
- Choose "Store a new secret"
- Select "Other type of secret"
- Create a secret with the following key-value pairs:
```
{
  "username": "your-bot-username",
  "password": "your-bot-password"
}
```
- Give the secret a name and description
- Complete the secret creation process
- Copy the secret's ARN
Configure Settings:
- Open the config.yaml file in the root directory of this repo
- Paste the secret ARN into the credentialsArn field

The Wickr bot setup depends on these resources being properly configured before deployment. If you deploy without completing these steps, the bot container won't initialize successfully.

Enable Bedrock Model Access

Before you can chat with the bot, you need to enable access to the model within your AWS account. This sample uses the Meta Llama 3 8B Instruct model.

Install Dependencies

From the root directory of the project, run:

npm install

Deployment

The bot is deployed as a container to AWS ECS Fargate. The deployment process:

Builds the Docker image
Pushes it to ECR
Updates the ECS service
Bot container (ECS Task) gets credentials from Secrets Manager
Bot container installs and starts the bedrock-bot integration

To deploy the project, run:

npm run cdk deploy

You should see output that looks similar to this when the deployment completes:

✅ WickrBedrockBotStack

✨ Deployment time: 223.78s

To confirm the Wickr Bot activated successfully, open the Wickr Admin console and navigate to User -> Bot Management. Your bot's status should now show as active.

Usage

To chat with the bot, open your AWS Wickr app and start a new Direct Message with your bot. If it doesn't automatically show up as a contact in your network, search for the bot username you created earlier. You can now chat with the bot, which will use the Amazon Bedrock foundation model to generate responses.

Cleanup

To clean up the environment and delete all provisioned infrastructure, run:

npm run cdk destroy --all --require-approval never

Security

See CONTRIBUTING for more information.

License

This library is licensed under the MIT-0 License. See the LICENSE file.

aws-samples/sample-bedrock-wickr-bot