API Tools

The MySQL MCP Server provides 14 tools for database operations, ML/GenAI, vector stores, and OCI Object Storage.

Database Connection Tools

list_all_connections

List all configured database connections and their provider modes. Parameters: None Returns: List of connection IDs with provider information (MySQL AI or HeatWave) Example:

"List all configured database connections"

Response:

[
  {
    "connection_id": "dev_db",
    "mode": "MySQL AI"
  },
  {
    "connection_id": "prod_heatwave",
    "mode": "HeatWave"
  }
]

execute_sql_tool_by_connection_id

Execute SQL statements on a specific database connection. Parameters:

connection_id (string, required): Database connection identifier
sql_script (string, required): SQL statement(s) to execute
params (object, optional): Parameters for parameterized queries

Returns: Query results or execution status Example:

"Execute 'SELECT * FROM employees LIMIT 10' on dev_db"

Supported Modes: MySQL AI, HeatWave

ML and GenAI Tools

ml_generate

Generate text using GenAI models. Parameters:

connection_id (string, required): Database connection identifier
question (string, required): Text generation prompt

Returns: Generated text response Example:

"Generate a summary of recent customer feedback using my_connection"

Supported Modes: MySQL AI, HeatWave

ragify_column

Create and populate a VECTOR column with embeddings from a text column. Parameters:

connection_id (string, required): Database connection identifier
table_name (string, required): Name of the table
input_column_name (string, required): Source text column
embedding_column_name (string, required): Target VECTOR column (will be created if it doesn’t exist)

Returns: Success status and number of rows processed Example:

"Add embeddings for 'body' column into 'embedding' column in docs table"

Supported Modes: MySQL AI, HeatWave

The ragify_column tool works identically on both MySQL AI and MySQL HeatWave connections.

heatwave_ask_help

Ask natural language questions about MySQL HeatWave AutoML features using NL2ML. Parameters:

connection_id (string, required): HeatWave database connection
question (string, required): Natural language question about HeatWave ML

Returns: Answer about HeatWave ML functionality Example:

"How do I train a model in HeatWave AutoML?"

Supported Modes: HeatWave only

This tool only works with MySQL HeatWave connections. It will return a clear error message if used with MySQL AI.

ask_nl_sql

Convert natural language questions into SQL queries and execute them automatically. Parameters:

connection_id (string, required): Database connection identifier
question (string, required): Natural language question

Returns: SQL query generated and execution results Example:

"What is the average delay incurred by flights?"
"Show me the top 5 customers by total order value"

Supported Modes: MySQL AI, HeatWave

retrieve_relevant_schema_information

Retrieve relevant schemas and tables (DDL) for a natural language question. Parameters:

connection_id (string, required): Database connection identifier
question (string, required): Natural language question

Returns: DDL statements for relevant tables and schemas Example:

"Which tables contain employee salary information?"

Supported Modes: MySQL AI, HeatWave

Vector Store Tools

list_vector_store_files_local

List available files in the secure_file_priv directory for local ingestion. Parameters:

connection_id (string, required): Database connection identifier

Returns: List of file paths available for loading Example:

"List local files for vector store ingestion"

Supported Modes: MySQL AI only

This tool does not work with MySQL HeatWave connections. Use load_vector_store_oci for HeatWave.

load_vector_store_local

Load documents from the local filesystem into a vector store. Parameters:

connection_id (string, required): Database connection identifier
file_path (string, required): Path to file in secure_file_priv directory

Returns: Loading status and document count Example:

"Load documents from /var/lib/mysql-files/manuals.pdf into vector store"

Supported Modes: MySQL AI only

load_vector_store_oci

Load documents from OCI Object Storage into a vector store. Parameters:

connection_id (string, required): HeatWave database connection
namespace (string, required): OCI Object Storage namespace
bucket_name (string, required): Bucket containing documents
document_prefix (string, required): Prefix filter for objects to load

Returns: Loading status and document count Example:

"Load all documents with prefix 'manuals/' from bucket 'company-docs' into vector store"

Supported Modes: HeatWave only

This tool only works with MySQL HeatWave (OCI) connections. For MySQL AI, use load_vector_store_local.

ask_ml_rag_vector_store

Perform RAG (Retrieval-Augmented Generation) query on the default vector store. Parameters:

connection_id (string, required): Database connection identifier
question (string, required): Natural language question
context_size (integer, optional): Number of context chunks to retrieve

Returns: Generated answer based on vector store contents Example:

"Ask ml_rag: Show me refund policy from the vector store"
"What does the documentation say about installation requirements?"

Supported Modes: MySQL AI, HeatWave

ask_ml_rag_innodb

Perform RAG query restricted to InnoDB tables using specified segment and embedding columns. Parameters:

connection_id (string, required): Database connection identifier
question (string, required): Natural language question
segment_col (string, required): Column containing text segments
embedding_col (string, required): Column containing vector embeddings
context_size (integer, optional): Number of context chunks to retrieve

Returns: Generated answer based on InnoDB table contents Example:

"Search my product_docs table for information about warranty policies"

Supported Modes: MySQL AI, HeatWave

On MySQL HeatWave, this tool loads the InnoDB table into the HeatWave cluster for maximum performance. On MySQL AI, processing occurs within the database service instance.

ask_ml_rag_vector_store (skip_generate mode)

Retrieve segments from the default vector store without text generation. Parameters:

connection_id (string, required): Database connection identifier
question (string, required): Natural language question
context_size (integer, optional): Number of segments to retrieve

Returns: Raw text segments without LLM generation Example:

"Retrieve relevant segments about installation from vector store"

Supported Modes: MySQL AI, HeatWave

OCI Object Storage Tools

These tools require a valid OCI configuration file (~/.oci/config). See Configuration for setup details.

list_all_compartments

List all compartments in your OCI tenancy. Parameters: None Returns: List of compartment names and OCIDs Example:

"List all compartments in my tenancy"

object_storage_list_buckets

List buckets in a specific compartment. Parameters:

compartment_name (string, optional): Compartment name
compartment_id (string, optional): Compartment OCID (alternative to name)

Returns: List of bucket names Example:

"Show all buckets in the development compartment"
"List buckets in ocid1.compartment.oc1..."

object_storage_list_objects

List objects in a bucket. Parameters:

namespace (string, required): OCI Object Storage namespace
bucket_name (string, required): Bucket name

Returns: List of object names and metadata Example:

"List objects in my 'docs-bucket'"

Tool Compatibility Matrix

Tool	MySQL AI	HeatWave	OCI Config Required
list_all_connections	✅	✅	❌
execute_sql_tool_by_connection_id	✅	✅	❌
ml_generate	✅	✅	❌
ragify_column	✅	✅	❌
heatwave_ask_help	❌	✅	❌
ask_nl_sql	✅	✅	❌
retrieve_relevant_schema_information	✅	✅	❌
list_vector_store_files_local	✅	❌	❌
load_vector_store_local	✅	❌	❌
load_vector_store_oci	❌	✅	✅
ask_ml_rag_vector_store	✅	✅	❌
ask_ml_rag_innodb	✅	✅	❌
list_all_compartments	✅	✅	✅
object_storage_list_buckets	✅	✅	✅
object_storage_list_objects	✅	✅	✅

Error Handling

The server provides clear, actionable error messages:

Unsupported Connection Type: When a tool is used with an incompatible database mode
Missing OCI Config: When cloud features are used without OCI configuration
SQL Errors: Database errors with query context
Parameter Validation: Missing or invalid tool parameters

Overview

MySQL

Oracle Database

Database Connection Tools

list_all_connections

execute_sql_tool_by_connection_id

ML and GenAI Tools

ml_generate

ragify_column

heatwave_ask_help

ask_nl_sql

retrieve_relevant_schema_information

Vector Store Tools

list_vector_store_files_local

load_vector_store_local

load_vector_store_oci

ask_ml_rag_vector_store

ask_ml_rag_innodb

ask_ml_rag_vector_store (skip_generate mode)

OCI Object Storage Tools

list_all_compartments

object_storage_list_buckets

object_storage_list_objects

Tool Compatibility Matrix

Error Handling

Next Steps

Vector Stores

ML & GenAI

Build docs developers (and LLMs) love

Overview

MySQL

Oracle Database

​Database Connection Tools

​list_all_connections

​execute_sql_tool_by_connection_id

​ML and GenAI Tools

​ml_generate

​ragify_column

​heatwave_ask_help

​ask_nl_sql

​retrieve_relevant_schema_information

​Vector Store Tools

​list_vector_store_files_local

​load_vector_store_local

​load_vector_store_oci

​ask_ml_rag_vector_store

​ask_ml_rag_innodb

​ask_ml_rag_vector_store (skip_generate mode)

​OCI Object Storage Tools

​list_all_compartments

​object_storage_list_buckets

​object_storage_list_objects

​Tool Compatibility Matrix

​Error Handling

​Next Steps

Vector Stores

ML & GenAI

Build docs developers (and LLMs) love

Database Connection Tools

list_all_connections

execute_sql_tool_by_connection_id

ML and GenAI Tools

ml_generate

ragify_column

heatwave_ask_help

ask_nl_sql

retrieve_relevant_schema_information

Vector Store Tools

list_vector_store_files_local

load_vector_store_local

load_vector_store_oci

ask_ml_rag_vector_store

ask_ml_rag_innodb

ask_ml_rag_vector_store (skip_generate mode)

OCI Object Storage Tools

list_all_compartments

object_storage_list_buckets

object_storage_list_objects

Tool Compatibility Matrix

Error Handling

Next Steps