execute_query
ChatGPTFor HubSpot portal SQL database execute analytical SQL query with optional parameters. Tip: When using AVG() on decimal/money columns (e.g. amount), wrap with CAST to avoid arithmetic overflow: AVG(CAST(amount AS FLOAT)). Read-only: this tool connects using a read-only SQL Server role; INSERT/UPDATE/DELETE/DDL and any other write statement are rejected by the database itself, not just validated client-side.
execute_saved_query
ChatGPTFor HubSpot portal SQL database execute a query previously saved via save_query, by the tool name it was saved under (as returned by list_saved_queries)
fetch
ChatGPTFor HubSpot portal SQL database retrieve the full content of a table, business terminology entry, or saved query previously returned by search, by its id
get_data_statistics
ChatGPTFor HubSpot portal SQL database get record counts for table, date ranges when available, null percentages, cardinality of categorical fields and common value distributions when calculated
get_database_schema_subtree
ChatGPTFor HubSpot portal SQL database get tables related to the table requested, and the table itself. The requested table, and any related table small enough to add little size (few columns), get full business meaning descriptions, categories, and allowed values per column. A wide related table (pulled in only because it has a foreign key to/from the requested table) shows column names, data types, and foreign key links only, to keep the response size down - call this tool again with that table's own name to get its full column detail. Prefer the default MARKDOWN output_format over JSON - MARKDOWN is far more compact for the same data; only request JSON when you specifically need to parse the result programmatically. A wide, heavily-customized table (e.g. a HubSpot Contact object with hundreds of custom properties) can still return a large amount of content even as MARKDOWN; if you only need specific fields rather than the whole table, use search/fetch instead
get_full_database_schema
ChatGPTFor HubSpot portal SQL database. WARNING This method returns large content - every table in the database at once, each with full business meaning descriptions per column. Prefer get_database_schema_subtree for a specific table instead - it is far smaller. Use the default MARKDOWN output_format over JSON unless you specifically need to parse the result programmatically - MARKDOWN is far more compact for the same data. Get database tables and business meaning descriptions for each, column names with business meaning descriptions and data types for each, primary keys and foreign key relationships, sample values and allowed values for enum-like columns
get_object_relationships
ChatGPTFor HubSpot portal SQL database get table relations and cardinality
get_semantic_metadata
ChatGPTFor HubSpot portal SQL database get business terminology and common aggregation patterns, alongside with categorical fields allowed values and their meanings
get_tables_list
ChatGPTFor HubSpot portal SQL database get tables and business meaning descriptions for each, primary keys and foreign key relationships
list_saved_queries
ChatGPTFor HubSpot portal SQL database list your saved queries (pre-built selections and custom queries saved via save_query), with their descriptions, query names and parameters. Use execute_saved_query to run one by name
save_query
ChatGPTFor HubSpot portal SQL database save SQL query and its parameters into a new or existing saved query for future reuse. Each saved query can consist of several queries sharing parameters fully or partially. Retrieve saved queries with list_saved_queries, run them with execute_saved_query
search
ChatGPTFor HubSpot portal SQL database search across tables, business terminology (semantic metadata), and saved queries for a keyword or phrase. Returns matching documents by id and title, with a link to your customer portal when available. Use fetch to retrieve the full content of a result
set_semantic_metadata
ChatGPTFor HubSpot portal SQL database contribute to business terminology and common aggregation patterns, categorical fields allowed values and their meanings. Note: value_info is optional; when provided it must list ALL the allowed values for this element (a partial list replaces, not merges with, any previously saved list)
execute_query
Claudeexecute_saved_query
Claudefetch
Claudeget_data_statistics
Claudeget_database_schema_subtree
Claudeget_full_database_schema
Claudeget_object_relationships
Claudeget_semantic_metadata
Claudeget_tables_list
Claudelist_saved_queries
Claudesave_query
Claudesearch
Claudeset_semantic_metadata
Claude