Building Conversational Interface
For Your Services
Focus on what / how you can help your users, instead of how many turns.
Leran more

Lay Schema At Base

Since backend API only speaks schema, to fulfill users’ needs, we need to convert user utterance into the frame events defined on the service schemas.
So CUI design needs to start with specifying the schemas for all your services. With the native support for user-defined types, interface, and arrays, Framely can naturally express the schema for any modern services, and their related business data and logic model.

Stack Annotations

Once we have schema, the goal of conversation is to fill the slots needed by it. Framely follow a practical style of conversation that can be decomposed into many self contained steps, each controlled by some dialog annotations attached to schema. By configuring dialog annotations, you define declaratively what conversational experience should be under given context in order to connect user needs with service. Based these annotations, Framely dialog engine can easily handle out of order conversation, or even transition back and forth between topics.

Sprinkle Expressions

Converting unstructured natural text into structured semantics, or frame events, that can be consumed by the dialog engine needs some serious natural language understanding (NLU) expertise. But no worry, with our sample efficient, nonparametric dialog understanding models, you can launch the chatbot and fix the understanding issues with just handful of example expressions, so anyone can do it.

Declarative
Framely makes it painless to create conversational UIs. Configure dialog annotations for each conversational condition in your service chatbot, Framely will conduct correct conversation with user based on current conversational state regardless of how conversation flows here, since user can switch topic back and forth, the number of flows can be uncountable. This declarative way of defining CUI is not efficient and also very easy to debug.
Reusable Component
Build frame, or conversational components, that manage their own state, then compose them to make complex conversational UIs. Framely support all modern language construct, and components logic is compiled to Kotlin, you can easily pass rich data through your chatbot. You can also import existing CUI components and save even more effort.
separate concerns
The goal of the conversation (what do say, defined by the interaction logic and current state) and the style of it (how to say, decided by who is the audience) are generally considered as different art from, and hence require different expertise. Using function to interaction backend, Framely allow CUI designer to easily template different conversational styles for different channels, in different languages. 
Open Sourced Runtime
The runtime will be open sourced. This not only add transparency so that you can be sure that code does the right thing, but also allow you to only commit to a technology instead of a company, since you can always tailor it to your needs. Of course, you can deploy it anywhere you want, so that you can make sure the user data is handled securely.
No code platform
Framely platform allow anyone with solid understanding of the business logic to build conversational experience for their services, no coding skill needed. Furthermore, the platform will generate runtime code in kotlin for a seamless integration into standard devops pipeline. 
Channels and support
The runtime is designed to support any number of channels to increase the reach of your service chatbot. At same time, we believe no matter how well you design and implement your chatbot, there is always the need for human agent to take over conversation. Framely is also designed to work any live chat software. As runtime is open source, your contribution is welcome.
A declarative way to develop chatbot !
Make it easy for people to communicate with your business.
send a request