Overview

HyperFlow AI uses a strongly-typed value system for data flowing between nodes. This includes simple data types like text and numbers, but also rich types like imported files, knowledge segments, generated images & video, and so on.

Making them "strongly-typed" means the system knows exactly what kind of data is flowing between the nodes and can perform all kinds of important checks and conversions and housekeeping, when needed.

Understanding Port Connections

When setting up data-flow connections between nodes in HyperFlow:

• Output ports produce values of a specific Data Value type, defined in the node's reference document
• Input ports define a main Data Value type they will accept, but often will accept other, compatible

Data Value types and will convert those value types as needed to perform its function.

Core Data Types

In the following, the Inputs accept: part defines the Data Value types that are acceptable connections to an Input port defined to be of the section's type.

So, for example, an input port of type text will accept connections from an output port of type text, prompt, instruction, knowledge, url, number, counter, model, cssSelectors, choice, pathName, code, json, content, importSet

Text

• Type Code: text
• Description: Plain text string values
• Inputs accept: text, prompt, instruction, knowledge, url, number, counter, model, cssSelectors, choice, pathName, code, json, content, importSet
• Common Use: General text processing, prompts, messages

Number

• Type Code: number
• Description: Numeric values (integers or floats)
• Inputs accept: number, counter, boolean, branch, text (parseable as a number)