execute_queryLABSCO SUMMARY
Vertica, answerable in conversation.
The streaming query is what makes it usable against a real warehouse: you can ask something that returns a lot of rows without waiting on one enormous response.
What it is
A Python MCP server for Vertica, installable with uvx or run from a Docker image. It executes queries, streams large result sets rather than returning them in one block, describes tables, lists views and indexes, and loads data with COPY.
What you get
- SQL executed against a Vertica database
- Large result sets streamed back rather than returned in one block
- A table's structure read: columns, types and constraints
- Views listed
- Indexes listed
- Data bulk-loaded with COPY
Requirements
A reachable Vertica instance and its credentials — VERTICA_HOST, VERTICA_PORT, VERTICA_DATABASE, VERTICA_USER and VERTICA_PASSWORD — set either as environment variables or as command-line flags. Connection limit and SSL behaviour are configured the same way.
Setup effort
One command plus credentials — uvx mcp-vertica, then supply credentials
USE CASES
Getting your bearings in an unfamiliar schema
List what views exist, then read the structure of the tables behind the ones you care about.
Answering an ad-hoc question with SQL
Run the query and read the result in the conversation instead of opening a client.
Pulling a result too large for one response
Stream it back in pieces rather than having the call fall over on size.
Loading a file into a table
Use COPY, which is how Vertica expects bulk data to arrive, rather than row-by-row inserts.
Working out why a query is slow
Check which indexes exist on the table and what its columns actually are before you rewrite the SQL.
6 tools
Listed in project docs. Your client may show a different list.
stream_querycopy_dataget_table_structurelist_indexeslist_viewsBEFORE YOU INSTALL
What to know first
Tools you get
6 tools named in the project's documentation.
Requires
- uvx on PATH
- VERTICA_PASSWORD in your environment
- VERTICA_USER in your environment
Where it runs
Both documented clients run the server on your machine.
INSTALLATION
Connect your client
Server setup
uvx mcp-verticaMCP Clients
Claude Desktop
{
"mcpServers": {
"vertica": {
"command": "uvx",
"args": [
"mcp-vertica"
],
"env": {
"VERTICA_HOST": "localhost",
"VERTICA_PORT": "5433",
"VERTICA_DATABASE": "VMart",
"VERTICA_USER": "dbadmin",
"VERTICA_PASSWORD": "test_password",
"VERTICA_CONNECTION_LIMIT": "10",
"VERTICA_SSL": "false",
"VERTICA_SSL_REJECT_UNAUTHORIZED": "true"
}
}
}
}Any MCP client — uvx
{
"mcpServers": {
"vertica": {
"command": "uvx",
"args": [
"mcp-vertica",
"--host=localhost",
"--db-port=5433",
"--database=VMart",
"--user=dbadmin",
"--password=test_password",
"--connection-limit=10"
]
}
}
}