google / bottery
Bottery is a syntax, editor, and simulator for prototyping generative contextual conversations modeled as finite state machines.
Bottery takes inspiration from the Tracery open-source project for generative text (also by katecompton@ in a non-google capacity) and the Cheap Bots, Done Quick! bot-hosting platform, as well as open FSM-based storytelling tools like Twine.
Like Tracery, Bottery is a syntax that specifies the script of a conversation (a map) with JSON. Like Cheap Bots, Done Quick!, the BotteryStudio can take that JSON and run a simulation of that conversation in a nice JavaScript front-end, with helpful visualizations and editing ability.
The goal of Bottery is to help everyone, from designers to writers to coders, be able to write simple and engaging contextual conversational agents, and to test them out in a realistic interactive simulation, mimicking how they'd work on a "real" platform like DialogFlow.
Users in Tracery write grammars, JSON objects that recursively define how to generate some text, like the musings of a lost self-driving car or outer-space adventures. Tracery grammars are lists of symbol names (like "animal") and their expansion rules (like "emu, okapi, pangolin").
