Points

Points are the primary components of the Point Engine. They are containers of logic, activated (i.e. filled) when directly called or their inputs are met. Filled points run the logic they contain, which can include a value, such as a string of text, a set of agent instructions, or code to be run. The logic can then produce outputs which may be written to the token. Points are stored as JSON objects.

Types of points are:

• Invoke: trigger code points when inputs are met
• Code: run structured or unstructured code
• Value: store non-confidential values
• Key: store confidential values