This component is used to provide instructions to the LLM, it has three working modes:
-
The value is used as a prompt to the LLM.
-
The value is used as a system prompt and the input is used as a prompt to the LLM.
-
It’s read-only with the prompt built in (eg. “Make it Real” Prompt).
Unnamed inputs are concatenated to the value and used as a prompt to the llm. Named inputs are wrapped with an xml tag <name>value</name>
and used as a prompt to the llm.
This block can also take an image or frame as input in case the selected model supports it.
The LLM services are provided by pollinations.ai.