What is it
Conversational projects require easy prototyping, fast implementation and continuous iteration. With Botfront you can build context-aware conversation flows in minutes and accelerate time to market by an exponential factor.
Easy to use
|Our main goal is to lower the technical barrier to entry in conversational AI. Implementing context-aware conversations should be as easy as chatting.||Botfront uses Rasa, a powerful and production ready conversational AI library. Botfront exposes all Rasa functionalities and abstracts its complexity.||Botfront's intuitive CLI orchestrates all Botfront's services on your development machine.
Authoring, training, auto reloading actions code, it just works!
Builds conversation as if you were chatting
|Botfront provides a unique and natural conversation authoring experience. You can create complex dialog flows in minutes|
Train & evaluate NLU models
|Botfront comes with a complete NLU toolbox. You can tag vast amounts of data efficiently, train and evaluate models. Several evaluation methods are available depending on the development stage of your model|
Annotate incoming data
|Botfront is always connected to your agent and conversation data keeps flowing in. You can annotate this data and even use it as an evaluation set and check how this new data improves your model.|
|Botfront exposes all Rasa features and concepts and and makes them accessible at a higher level for faster development. You can export a Botfront project and use it with Rasa at any time.|
Botfront only requires a recent version of Docker. You can install the CLI with the following:
npm install -g botfront
Then just run
botfrontto get started.
Botfront is a Meteor app, so the first step is to install Meteor
Then clone this repo and install the dependencies:
git clone https://github.com/botfront/botfront cd botfront/botfront meteor npm install
Install the CLI from the source code:
# if you installed Botfront from npm uninstall it. npm uninstall -g botfront # Install the cli from the source code cd cli && npm link
Botfront needs to be connected to other services, especially Rasa. To do this, you need to create a regular project, and start Botfront with a dedicated configuration:
Create a Botfront project with
botfront init(somewhere else, not in the repo)
Start your project with
botfront up -e botfront. This will run all services except the Botfront app, since you are going to run it with Meteor locally
Go back to the botfront checkout
cd botfront/botfrontand run Botfront with
meteor npm run start:docker-compose.dev. Botfront will be available at http://localhost:3000 so open your browser and happy editing :smile_cat:
Some botfront cli commands that may help if you run into problems:
botfront init # create a new botfront project botfront logs # show the logs! botfront killall # stop all docker services botfront down # stop all botfront services botfront up # restart botfront botfront docs # open the docs in your browser
Note that these should be run from the same directory as your botfront project
We ❤️ contributions of all size and sorts. If you find a typo, if you want to improve a section of the documentation or if you want to help with a bug or a feature, here are the steps:
- Fork the repo and create a new branch, say
- Fix/improve the codebase
- Commit the changes. Commit message must follow the naming convention, say
fix(conversation builder): display story groups in alphabetical order
- Make a pull request. Pull request name must follow the naming convention. It can simply be one of your commit messages, just copy paste it, e.g.
fix(readme): improve the readability and move sections
- Submit your pull request and wait for all checks passed (up to an hour)
- Request reviews from one of the developers from our core team.
- Get a 👍 and PR gets merged.
Well done! Once a PR gets merged, here are the things happened next:
- all Docker images tagged with
branch-masterwill be automatically updated in an hour. You may check the status on the Actions tab.
- your contribution and commits will be included in our release note.
End to end tests are using the Cypress testing framework.
To manually run the Cypress tests, you need to have Botfront running in development mode. Some tests also require Rasa to be available.
Once you are at the root of the repo, you can enter the following.
cd botfront # if you want to open Cypress' graphical interface npx cypress open # If you want to run the whole suite in headless mode # This could take up to an hour depending on your computer npx cypress run # If you want to run a specific test npx cypress run --spec "cypress/integration/02_training/training.spec.js"
Commit messages naming convention
To help everyone with understanding the commit history of Botfront, we employ
commitlint to enforce the commit styles:
type is one of the following:
scope is optional, represents the module your commit working on.
subject explains the commit.
As an example, a commit that improved the documentation:
docs(conversation builder): update slots manager screenshot.