Skip to content
Go to file
This branch is 4 commits ahead, 616 commits behind aws-samples:master.

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Sample Amazon Lex Web Interface


This repository provides a set of AWS CloudFormation templates to automatically build and deploy a sample Amazon Lex web interface.

Web UI Application

The sample chatbot web user interface can be used to interact with an Amazon Lex bot using text or voice with a webRTC capable browser.

The interface interacts with AWS services directly from the browser. Here is a diagram of how the application works:

For details about the web app, see the accompanying README file and the source under the lex-web-ui directory.

Here is a screenshot of it:


To deploy a CloudFormation stack with a working demo of the application, follow the steps below:

  1. Click the following CloudFormation button to launch your own copy of the sample application stack in the us-east-1 (N. Virginia) AWS region: cloudformation-launch-stack. You can accept the defaults in the CloudFormation Create Stack Wizard up until the last step. At the last step, when prompted to create the stack, select the checkmark that says: "I acknowledge that AWS CloudFormation might create IAM resources with custom names". It takes about 10 minutes for the CloudFormation stacks to got to CREATE_COMPLETE status.
  2. Once the status of all the CloudFormation stacks is CREATE_COMPLETE, click on the PipelineUrl link in the output section of the master stack. This will take you to the CodePipeline console. You can monitor the progress of the deployment pipeline from there. It takes about 10 minutes to build and deploy the application.
  3. Once the pipeline has deployed successfully, go back to the output section of the master CloudFormation stack and click on the ParentPageUrl link. You can also browse to the WebAppUrl link. Those links will take you to the sample application running as an embedded iframe or as a stand-alone web application respectively.

CloudFormation Stack


Here is a diagram of the CloudFormation stack created by this project:

CloudFormation Resources

The CloudFormation stack creates the following resources in your AWS account:

  • A Amazon Lex bot. You can optionally pass the bot name of an existing one to avoid creating a new one.
  • A Cognito Identity Pool used to pass temporary AWS credentials to the web app. You can optionally pass the ID of an existing Cognito Identity Pool to avoid creating a new one.
  • A CodeCommit repository loaded with the source code in this project
  • A continuous delivery pipeline using CodePipeline and CodeBuild. The pipeline automatically builds and deploys changes to the app committed to the CodeCommit repo.
  • An S3 bucket to store build artifacts
  • Two S3 buckets to host the web application (parent and iframe). The pipeline deploys to this bucket.
  • Lambda functions used as CloudFormation Custom Resources to facilitate custom provisioning logic
  • CloudWatch Logs groups automatically created to log the output of Lambda the functions
  • Associated IAM roles for all of the above

CloudFormation Templates

The CloudFormation launch button above launches a master stack that in turn creates various nested stacks. The following table lists the CloudFormation templates used to create these stacks:

Template Description
templates/master.yaml This is the master template used to deploy all the stacks. It uses nested sub-templates to include the ones listed below.
templates/lexbot.yaml Lex bot and associated resources (i.e. intents and slot types).
templates/cognito.yaml Cognito Identity Pool and IAM role for unauthenticated identity access.
templates/coderepo.yaml CodeCommit repo dynamically initialized with the files in this repo using CodeBuild and a custom resource.
templates/pipeline.yaml Continuous deployment pipeline of the Lex Web UI Application using CodePipeline and CodeBuild. The pipeline takes the source from CodeCommit, builds the Lex web UI application using CodeBuild and deploys the app to an S3 bucket.


When launching the stack, you will see a list of available parameters and a brief explanation of each one. You can take the default values of most of the CloudFormation parameters. The parameters that you may need to modify are:

  • BotName: Name of pre-existing Lex bot. This is an optional parameter. If left empty, a sample bot will be created based on the OrderFlowers bot in the Lex Getting Started documentation.
  • CognitoIdentityPoolId: Id of an existing Cognito Identity Pool. This is an optional parameter. If left empty, a Cognito Identity Pool will be automatically created. The pool ID is used by the web ui to get AWS credentials for making calls to Lex and Polly.
  • ParentOrigin: Origin of the parent window. Only needed if you wish to embed the web app into an existing site using an iframe. The origin is used to control which sites can communicate with the iframe


Once the CloudFormation stack is successfully launched, the status of all nested stacks will be in the CREATE_COMPLETE green status. At this point, the master stack will reference the resources in the output section. Here is a list of the output variables:

  • PipelineUrl: Link to CodePipeline in the AWS console. After the stack is successfully launched, the pipeline automatically starts the build and deployment process. You can click on this link to monitor the pipeline.
  • CodeCommitRepoUrl: CodeCommit repository clone URL. You can clone the repo using this link and push changes to it to have the pipeline build and deploy the web app
  • WebAppUrl: URL of the web app running on a full page. The web app will be available once the pipeline has completed deploying
  • ParentPageUrl: URL of the web app running in an iframe. This is an optional output that is returned only when the stack creates the sample when page. It is not returned if an existing origin is passed as a parameter to the stack during creation.
  • CognitoIdentityPoolId: Pool ID of the Cognito Identity Pool created by the stack. This is an optional output that is returned only when the stack creates a Cognito Identity Pool. It is not returned if an existing pool ID was passed as a parameter to the stack during creation.

Deployment Pipeline

When the stacks have completed launching, you can see the status of the pipeline as it builds and deploys the application. The link to the pipeline in the AWS console can be found in the PipelineUrl output variable of the master stack.

Once the pipeline successfully finishes deploying, you should be able to browse to the web app. The web app URL can be found in the WebAppUrl output variable.

The source of this project is automatically forked into a CodeCommit repository created by the CloudFormation stack. Any changes pushed to the master branch of this forked repo will automatically kick off the pipeline which runs a CodeBuild job to build and deploy a new version of the web app. You will need to setup CodeCommit to push changes to this repo. You can obtain the CodeCommit git clone URL from the CodeCommitRepoUrl output variable of the master stack.

Here is a diagram of the deployment pipeline:

Directory Structure

This project contains the following main directories:

 |__ build                 # Makefile used for uploading the project sources into S3
 |__ lex-web-ui            # sample Lex web ui application source
 |__ templates             # cloudformation templates and related lambda functions
    |__ custom-resources   # lambda functions used in cfn custom resources

How do I ...?

Use or deploy my own bot?

The BotName CloudFormation parameter can be used to point the stack to an existing bot. In the application, you can also change the configuration files or pass parameters to it (see the application README file for details).

If you want to make changes to the sample bot deployed by the stack, you can edit the bot-definition.json file. This file is used by the which is run in Lambda by a CloudFormation Custom Resource in the bot stack created by the lexbot.yaml template. The bot definition is in a JSON file that contains all the resources associated with the bot including intents and slot types.

The script can be also used as a stand-alone shell script. It allows to export existing bots (including associated resources like intents and slot types) into a JSON file. The same script can be used to import a bot definition into an account or to recursively delete a bot and associated resources. Here is the script usage:

$ python  -h
usage: [-h] [-i [file] | -e [botname] | -d botname]

Lex bot manager. Import, export or delete a Lex bot. Used to
import/export/delete Lex bots and associated resources (i.e. intents, slot

optional arguments:
  -h, --help            show this help message and exit
  -i [file], --import [file]
                        Import bot definition from file into account. Defaults
                        to: bot-definition.json
  -e [botname], --export [botname]
                        Export bot definition as JSON to stdout Defaults to
                        reading the botname from the definition file: bot-
  -d botname, --delete botname
                        Deletes the bot passed as argument and its associated

Delete the CloudFormation stacks?

The resources created by this stack can be easily removed from your account by deleting the master CloudFormation stack. The master stack is the one that was first created using the "Launch Stack" button. By deleting this stack, the rest of the sub-stacks and resources will be deleted with the exception of the CloudWatch Logs groups created by the stack (these are retained for troubleshooting purposes).

The S3 buckets created by the stacks are deleted by default. If you wish to retain the data in these buckets, you should set the CleanupBuckets parameter to false in the master stack.

Deploy Using My Own Bootstrap S3 Bucket

The source used to bootstrap the CodeCommit repo created by CloudFormation is dynamically downloaded from a predefined S3 bucket. If you want to use your own S3 bucket, this project provides a Makefile under the build directory to facilitate uploading the bootstrap artifacts. Follow these steps:

  1. Create your own S3 bucket (you might want to enable bucket versioning on this bucket).
  2. Modify the master.yaml template to point to your bucket. The bucket and path are configured by the BootstrapBucket and BootstrapPrefix variables under the Mappings section of the template.
  3. Modify the variables in the local build environment file: build/config.env. These variables control the build environment and web application deployment. In specific, you should modify the following variables:
    • BOOTSTRAP_BUCKET_PATH: point it to your own bucket and prefix merged together as the path to the artifacts (same as step 2)
  4. Upload the files to your S3 bucket using make upload. The build directory under the root of the repo contains a Makefile that can be used to build the artifacts and upload the files to your S3 bucket. It uses the aws cli to copy the files to S3. To upload the files to your s3 bucket, issue the following commands (from the root of the repository):
cd build
make upload # requires a properly configured aws cli installation
You can’t perform that action at this time.