ย Github
Authoring conversations
Monitoring & Analyzing
Configuring channels
On screen guidance
Proactive conversations
On your local machineRequirementsInstall the Botfront CLICreate a new project folderStart BotfrontCreate your admin accountCreate your projectConnect Rasa to your new projectSecure rasa access (optional)On servers or clustersUsing the CLIMigration guide


Botfront is multi-project since Botfront 1.x. This feature is more useful on a hosted environment than on a local machine where it's more convenient to have one project per folder. Hence this process might seem a bit convoluted


Botfront has a few technical requirements. See them here.

  • Linux or macOS (Windows may work but is not officially supported yet, help is welcome)
  • Chrome (some issues with other browsers)
  • A recent version of Docker
  • A recent version of Node.js

Install the Botfront CLI

npm install -g botfront

Create a new project folder

botfront init

Botfront will prompt you for a folder name if the current directory is not empty

Start Botfront

botfront up

Botfront will download Docker images and then start. Botfront will open in your browser.

Create your admin account

Just fill the usual details and save.

Create your project

Do NOT use the chitchat project

Above the project list click Add project

And enter project details. Namespace doesnโ€™t matter in local installations, and submit

Connect Rasa to your new project

Copy your project id

Then, in your terminal (from the project folder):

botfront set-project 3PEj5txDenAwG5brc # or whatever your project ID is

Congrats! Your project is now ready!

Secure rasa access (optional)

You might want to secure the communication between botfront and rasa-for-botfront.

  • In the instance settings there is a token, you can edit it or generate new ones as you feel.

  • Add this token in env part of the .botfront/botfront.yml under the key auth_token

auth_token: 5b576c56-45fd-4d30-9a07-8a701e48f2f3
bf_project_id: bf
bf_url: 'http://host.docker.internal:3000/graphql'
root_url: 'http://localhost:8888'
augmentation_factor: 50
mongo_url: 'mongodb://mongo:27017/bf'
  • Restart Botfront (botfront down and then botfront up)

The Rasa API is now protected !