Hi, folks.
I developed AI-templates for Obsidian Templater for new knowledge development (yeah, yet another templates, I know… :)). The valuable features:
-
ready-to-use templates (with default settings)
-
structured LLM prompting
-
maximal efficiency of LLM prompting via aspect management
-
flexible LLM output management.
The simplest way to generate new information in Obsidian appears to be as follows: while working in a specific .md document, the user selects text in the editor to which they apply a specific Templater template. This template corresponds to a specifically configured AI function and the necessary aspects for the particular task. The AI function is executed, and the result is inserted into the current document immediately following the selected text (for further editing or new generation). In this scenario, the following data elements can be used as input for the AI function:
-
The selected text to which the AI function should be applied.
-
The file name, which defines the “topic” or knowledge domain in which generation is required.
-
The full content of the current file.
-
The values of the aspects for the given task, defined in the Templater template code.
By selecting the necessary information and generating new information through customizable templates, one can develop the “knowledge” described in the current document.
The templates code:
https:// github .com /ady1981/obsidian-templater-core-kbt
I hope to get some useful feedback.