MCP App Store
by KeenEthics

Overview

NLSQL turns a company database into a data source that people can query in plain English. This plugin sets that data source up for you in the chat, so you don't have to fill in the NLSQL configuration forms by hand. What it does: you paste your database schema (the output of a schema-only dump command the plugin gives you for MySQL, PostgreSQL, MSSQL, BigQuery, Snowflake, Amazon Redshift, SQLite or CQL). The plugin reads the tables, columns and joins, lets you choose what to include, and captures your KPIs, top-N rankings, filters and calculated metrics from plain-language descriptions. It then publishes the configuration to your NLSQL account. Afterwards you can add the known values of category columns to make answers more accurate. Who it is for: data, BI and engineering teams setting up NLSQL for their organisation. An NLSQL account is required; you sign in with it the first time you use the plugin. Limitations: the plugin never connects to your database and never asks for credentials or table data. It works only from the schema text you provide. It configures NLSQL but does not run queries or answer questions about your data; that happens in NLSQL's own apps once the data source is published. One data source is set up at a time per account.

Tools

get_column_values

ChatGPT
Add distinct column values to the user's published DataSource.

nlsql_config_wizard

ChatGPT
Guides the user through the entier NLSQL configuration process, building their DataSource object. Consists of the following steps: 1. DataSource name 2. Database type 3. Database schema (shell command + raw output) 4. Schema interpretation (llm extracts tables, columns, joins) 4.5 Schema confirmation 5. Table selection 5.5 Column selection (per table) 5.6 Joins confirmation 6. Default date column detection + selection 7. Arguments (KPI aggregations) 8. Top arguments (top-N rankings) 9. Filters (static column filters) 10. Complex arguments (cross-column calculations) 11. Column arguments (column-value named arguments) 12. Verify the completed DataSource configuration State is persisted automatically. Calls with no data will resume from where the user left off. The'data' argument is used to pass the user's response for the current step.

publish_column_values

ChatGPT
Publish the user's distinct column values to NLSQL

publish_datasource

ChatGPT
Publish the user's complete DataSource

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps