Botfront.
 Github
Authoring conversations
Monitoring & Analyzing
Configuring channels
On screen guidance
Proactive conversations
Installation
On your local machineOn servers or clustersIntegrationAvailable webhooksUsing the CLIMigration guide
Developers
Administration
Deployment

Available in: Enterprise Edition

Webhooks

Some Botfront features implies being able to communicate with the underlying IT infrastructure. Since every infrastructure is specific, some feature may require custom implementation to be available in Botfront. For example, restarting a Rasa instance, uploading medias, or deploying are very specific on how your IT infrastructure is organized.

Webhooks allow a deeper integration in your IT environment.

You can define webhooks at the platform lever in Admin -> Settings -> Webhooks.

Available webhooks

Restart Rasa

If set, this webhook will be invoked when changing endpoints, credentials, or performing any action requiring Rasa to restart.

Expected request body

{
"projectId": string,
"environment" : string,
}

Expected response

This route does not return a an object. Botfront only expects the http status code.

Upload Image

Implementing this webhook enables image upload in Botfront.

Expected request body

{
"projectId": string,
"data": string, // image encoded in base64
"mimeType": string,
"language": string,
"responseId": string // template name followed by unix timestamp, e.g. utter_get_started_1588107073256
}

Expected response

{
"uri": string // the publicly accessible URI for the asset just received
}

Delete Image

The webhook will be invoked by Botfront when an image URL is no longer used in a response. The request will be sent whether or not the resource actually exists on the server.

Expected request body

{
"projectId": string,
"uri": string // the publicly accessible URI for the asset to be deleted
}

Expected response

Botfront expects a 204 success response, and a 404 if ressource was not found. Any other response may signal an error to the user in the future.

Deploy

If set, and if your project has environments defined, the deploy menu will be available next to the training button. When training, the webhook will be invoked with the development environment. When deploying, the webhook will be invoked with the production environment.

Request

Botfront will send a request to the api with the following parameters.

{
"projectId" : string,
"namespace" : string,
"environment" : string, // "development" or "production"
"data" : string, // base64 conversion of the model archive
"mimeType" string,
}

Expected response

Botfront except a 200 success response, other response are treated as errors. You can and a message property in the error response body that will be displayed to the user.

Report Crash

If set, browser crashes traces will sent to this webhook.

Request

Botfront will forward the following to the webhook:

{
"version": string,
"path": string,
"error": string,
"trace": string
}

Expected response

Success-like responses (e.g. 200) are treated as successful.

🖊️ Edit this page on Github