IBM watsonx MCP Server

create_prompt

Create a new prompt in watsonx

generate_chat

Use this for multi-turn conversational AI applications. Generate chat completions using a watsonx chat model

generate_embeddings

Useful for similarity search, clustering, and semantic analysis. Generate vector embeddings for input texts

generate_text

Use this for single-turn text generation tasks like content creation, summarization, or analysis. Generate text using a watsonx foundation model

get_model_details

Get detailed specifications for a specific foundation model

get_tuning_status

Get the status of a prompt tuning job

list_models

ai, including model IDs, families, capabilities, and lifecycle states. List available foundation models in watsonx

list_projects

List watsonx projects in your account

list_prompts

List saved prompts in the watsonx project

start_model_tuning

Requires a URL pointing to the training data in cloud storage. Start a prompt tuning job for a foundation model