About this project

TM Kamishibai is a participatory AI storytelling framework where participants move a story forward by performing specific poses in front of a camera. The system leverages Teachable Machine (TM) for pose recognition and TurboWarp for authoring and playback, allowing stories to be deployed as Web apps or SB3 projects. Key capabilities include: - Story Authoring: Support for describing scenes, dialogue, motion, and branching narratives using a text-based DSL (versions 3.1/3.2) or YAML (version 4.0). - Interactive Controls: Ability to trigger story progression via pose recognition, keyboard input, and stage touches. - Tooling: A CLI for automating DSL validation, live previews, and building self-contained SB3 files. - Integration: JavaScript APIs and TurboWarp blocks to invoke core actions. The project provides a runtime, a set of developer tools for CI/CD and reproducible builds, and a comprehensive documentation site for authors and workshop organizers.