Application of LLM function call from MCP document.
IMcpLlmApplication is an interface representing a collection of
LLM function calling schemas composed from the
MCP (Model Context Protocol) document. It contains
failed functions, and adjusted
options during the IMcpLlmApplication
construction.
About each function of MCP server, there can be errors during
the composition, if the target model does not support the
function's IMcpLlmFunction.parameters type. For example,
Google Gemini model does not support union type, so that the function
containing the union type would be placed into the errors list
instead of functions.
Also, each function has its own IMcpLlmFunction.validate
function for correcting AI agent's mistakes, and this is the reason why
@samchon/openapi recommends not to use the
mcp_servers
property of LLM API directly, but to use the function calling feature
instead.
Application of LLM function call from MCP document.
IMcpLlmApplicationis an interface representing a collection of LLM function calling schemas composed from the MCP (Model Context Protocol) document. It contains failed functions, and adjusted options during theIMcpLlmApplicationconstruction.About each function of MCP server, there can be errors during the composition, if the target model does not support the function's IMcpLlmFunction.parameters type. For example, Google Gemini model does not support union type, so that the function containing the union type would be placed into the errors list instead of functions.
Also, each function has its own IMcpLlmFunction.validate function for correcting AI agent's mistakes, and this is the reason why
@samchon/openapirecommends not to use themcp_serversproperty of LLM API directly, but to use the function calling feature instead.Author
Jeongho Nam - https://github.com/samchon
Author
Byeongjin Oh - https://github.com/sunrabbit123