Understand how prompts are structured in Athina
There are a few important concepts to understand:
prompt_slug: This is the name of your prompt. It must be unique for your workspace. You can see the prompt slug in the URL and in the sidebar in the playground.
prompt_template: This is a versioned prompt template. Every prompt_template
must belong to a prompt_slug
. Draft prompts are also stored as prompt_template
but without an associated version
.
version (integer): This is the current version of the prompt template. Versions are stored as auto-incrementing integers. When you commit a prompt, a new version is created.
is_live: When you deploy a prompt_template
, it is marked as live
. The live
prompt is the one that is used when you run the prompt via API (unless overridden with an explicit version).
prompt_execution: This refers to an individual execution (or run) of a prompt_template
in the playground.
A saved prompt_template
in Athina looks like this:
Understand how prompts are structured in Athina
There are a few important concepts to understand:
prompt_slug: This is the name of your prompt. It must be unique for your workspace. You can see the prompt slug in the URL and in the sidebar in the playground.
prompt_template: This is a versioned prompt template. Every prompt_template
must belong to a prompt_slug
. Draft prompts are also stored as prompt_template
but without an associated version
.
version (integer): This is the current version of the prompt template. Versions are stored as auto-incrementing integers. When you commit a prompt, a new version is created.
is_live: When you deploy a prompt_template
, it is marked as live
. The live
prompt is the one that is used when you run the prompt via API (unless overridden with an explicit version).
prompt_execution: This refers to an individual execution (or run) of a prompt_template
in the playground.
A saved prompt_template
in Athina looks like this: