execute_sqlLABSCO SUMMARY
Redshift, with the schema already on the table.
The assistant browses schemas and DDL as resources before writing a query, then reads the execution plan instead of discovering the cost by running it.
What it is
A Python server for Amazon Redshift. Three tools do the work — run SQL, read a plan, analyze a table — while schemas, table lists, DDL and table statistics are exposed as MCP resources the client can browse before it writes anything.
What you get
- execute_sql runs a query against the cluster
- get_execution_plan returns the plan for a query before you commit to running it
- analyze_table collects statistics on a table
- Resources cover the reading: rs:///schemas lists schemas, rs:///{schema}/tables lists tables, and each table's DDL and statistics have their own rs:/// address
Requirements
Python 3.13 or higher, uv, and a reachable Redshift cluster. Connection details come from environment variables: RS_HOST, RS_PORT, RS_USER, RS_PASSWORD, RS_DATABASE and an optional RS_SCHEMA.
Setup effort
Build from source — git clone https://github.com/Moonlight-CL/redshift-mcp-server.git
USE CASES
01
Answering a question against the warehouse
The query runs on the cluster and the rows come back into the conversation, with the schema already in view.
02
Checking what a query will cost before running it
The execution plan comes back first, which on a warehouse is the difference between a second and a coffee break.
03
Fixing a query the planner is getting wrong
Refreshing a table's statistics and reading the plan again shows whether the estimate was the problem.
3 tools
Listed in project docs. Your client may show a different list.
analyze_tableget_execution_planBEFORE YOU INSTALL
What to know first
Tools you get
3 tools named in the project's documentation.
Requires
- uv on PATH
- Python 3.13 or later
- An Amazon Redshift cluster you can reach
- Redshift login for it: RS_HOST, RS_PORT, RS_USER, RS_PASSWORD and RS_DATABASE in the client configuration
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
git clone https://github.com/Moonlight-CL/redshift-mcp-server.git
cd redshift-mcp-server
uv syncAdd the configuration under MCP Clients to your client, with the full path of src/redshift_mcp_server inside your clone and your cluster's connection details
MCP Clients
Any MCP client
ANY MCP CLIENT · JSON
{
"mcpServers": {
"redshift": {
"command": "uv",
"args": [
"--directory",
"<YOUR_PATH_TO_REDSHIFT_MCP_SERVER>/src/redshift_mcp_server",
"run",
"server.py"
],
"env": {
"RS_HOST": "<YOUR_REDSHIFT_CLUSTER>.region.redshift.amazonaws.com",
"RS_PORT": "5439",
"RS_USER": "<YOUR_RS_USER>",
"RS_PASSWORD": "<YOUR_RS_PASSWORD>",
"RS_DATABASE": "<YOUR_RS_DATABASE>",
"RS_SCHEMA": "<YOUR_RS_SCHEMA>"
}
}
}
}<YOUR_PATH_TO_REDSHIFT_MCP_SERVER> is the full path of the redshift-mcp-server folder the clone step creates — the README's relative src/redshift_mcp_server only resolves if the client happens to start in that folder. Fill the RS_ values with your cluster's details; RS_SCHEMA is optional and defaults to public.
