Coordination Editor

The Coordination Editor is designed around the core concepts of Nodes and Canvases. A Node represents an individual, functional unit of data, complete with a unique ID, title, and associated content. These Nodes serve as building blocks within a Canvas, which is essentially a collection of interconnected Nodes.

There are currently a selection of different Nodes types to work with in the platform. They serve as building blocks that can be stacked together to create a Method

• Default: Default Node with a Node Title and Node Page

• Loop: Instructional Node that tells the AI to loop through the preceding Canvas/Nodes

• Output: States where the AI should output information

• Prompt: Instructions for the AI to follow

• Response (combined): Output into a Canvas

• Response (one node): Outputs from a loop into a Canvas with one node response per node input of the loop

• Response (many nodes): Outputs from a loop into a Canvas with many nodes’ response per node input of the loop

• Paper Finder: Calls the Semantic Scholar API

Nodes and Canvases work together via a contextual zoom. In the below figure we click on the TCE Innovation Strategies Node, which opens a Canvas of Nodes within. This "zoom in" feature lets users explore and create content at different levels of detail, from broad overviews to specific information. When working with the AI, users can choose how deep they want to go, deciding how many times they "zoom in" for more detail.

Viewing at different levels of depth allows more details to be added for context for the AI while allowing the information to be summarized depending on the needs of the topic. A breadcrumb trail allows for quick navigation along the hierarchy.

Users can bring in Nodes from any Canvas to outline a topic or idea at any level of depth. These Nodes are all embedded meaning that an update to the Node in one location is automatically updated in all other locations. This allows users to bring together and use outputs from different topics and ideas into other ones. And it creates the use of a type of dashboard or the idea of embedding outputs of multiple projects into one Canvas.

Each Node is enriched with metadata (future feature), which enables powerful features such as traceability and provenance. These capabilities ensure that users can track the history and evolution of information, addressing common challenges like cluttered and outdated content, and version control. By embedding metadata into every Node, the platform allows the AI to coordinate information, alleviating the burden of manual organization. This structure ensures that users always have access to the most accurate, relevant data without the typical "garbage in, garbage out" issues that plague traditional systems. The intuitive design of the platform makes it easy for users to get started and continue making meaningful progress without unnecessary complexity.