Data Platform MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MCP_PORT | No | Puerto local publicado. | 8000 |
| IMAGE_TAG | No | Etiqueta local de la imagen construida. | 0.1.0 |
| LOG_LEVEL | No | Nivel de log de Uvicorn. | info |
| MCP_BIND_ADDRESS | No | Interfaz del anfitrión donde se publica el puerto. | 127.0.0.1 |
| AI_PLATFORM_NETWORK | No | Red externa compartida con Open WebUI. | ai-platform |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| logging | {} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| extensions | {
"io.modelcontextprotocol/ui": {}
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| hello_worldA | Return a deterministic greeting used to verify MCP connectivity. |
| health_checkA | Report MCP process liveness and server/contract versions. |
| list_connectionsA | List configured connections and adapter capabilities without secrets. |
| get_connection_capabilitiesB | Return one connection and its adapter capability matrix without secrets. |
| test_connectionB | Test an enabled connection using a bounded read-only operation. |
| refresh_schema_cacheA | Refresh one connection, or every enabled connection when omitted. |
| get_schema_cache_statusA | Return freshness and last refresh state for one or all connections. |
| search_catalogB | Search cached tables, columns, descriptions and relationships. |
| list_schemasA | List visible schemas from the selected connection's cached snapshot. |
| list_tablesA | List cached tables, optionally restricted to one schema. |
| describe_tableB | Return columns, keys and comments for one cached table. |
| list_relationshipsC | List foreign keys whose source or target matches the optional filters. |
| validate_sqlA | Parse and classify SQL for the selected connection without executing it. |
| execute_read_queryA | Execute one validated SELECT under row, byte, timeout and concurrency limits. |
| explain_queryC | Return a JSON PostgreSQL plan for a SELECT without using ANALYZE. |
| generate_sqlA | Generate one validated SELECT from a natural-language question, without executing it. |
| generate_and_execute_queryC | Generate SQL and, only if it is executable, run it under full revalidation. |
| generate_reportC | Generate SQL, execute it under full revalidation and export it inline. |
| list_proceduresB | List cached procedures and functions, optionally restricted to one schema. |
| list_triggersA | List cached triggers, optionally restricted to schema and/or table. |
| explain_database_objectB | Explain one cached procedure or trigger from its real SQL definition. |
| search_documentsC | Retrieve the most relevant cached document chunks for a query. |
| list_indexed_documentsB | List indexed documents, optionally restricted to connection and/or domain. |
| refresh_document_indexB | Reindex a single file, or scan the documents directory completely. |
| delete_indexed_documentB | Remove one document from the vector store and its cached metadata. |
| list_mongo_collectionsA | List visible, non-system collections for a MongoDB connection. |
| validate_mongo_queryA | Validate a find filter or aggregation pipeline without executing it. |
| execute_mongo_findB | Execute one validated find() under row, byte, timeout and concurrency limits. |
| execute_mongo_aggregateB | Execute one validated aggregation pipeline under the same limits as find(). |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 29 tools
Several tools overlap in purpose, particularly hello_world vs health_check, and the family of generate_and_execute_query, generate_report, generate_sql, and execute_read_query. The descriptions help clarify boundaries, but the large number of similar actions creates potential for misselection.
Most tools follow a consistent verb_noun pattern (list_schemas, execute_read_query, refresh_schema_cache). Minor deviations like hello_world, health_check, and generate_and_execute_query break the pattern, but the convention is largely readable and predictable.
At 29 tools, the server exceeds the typical well-scoped range. There is notable redundancy (hello_world and health_check both check connectivity; separate SQL and Mongo execute tools could be unified) and the count feels heavy for an agent to navigate.
The tool surface covers schema discovery, SQL validation/execution/explanation, Mongo querying, document indexing, and connection management. Missing write operations for SQL/Mongo appear intentional for a read-only platform, so the gaps are minor.