execute_query
Execute SQL queries against the Vertica database. Query restrictions depend on configuration.
LABSCO SUMMARY
A result set too large to return in one response stops being a reason to leave the conversation.
stream_query is what separates this from a plain query wrapper: batchSize and maxRows keep a wide table from arriving as a single payload. On the introspection side, list_indexes reports Vertica projections, which is the structure that actually governs performance in that engine. Whether execute_query can write is a configuration question rather than a property of the tool, so a read-only deployment is something you set up rather than something you are given.
What it is
A Vertica client covering query execution and schema introspection.
What you get
- SQL execution with bound parameters, under whatever query restrictions the deployment is configured with.
- Batched streaming of large result sets, with the batch size and a maximum row count in your hands.
- Tables and views in a schema with their metadata, and the definitions behind the views.
- A table's columns, data types and constraints, plus its projections with column and uniqueness information.
Requirements
A reachable Vertica database and its connection settings; what a query is permitted to do depends on how the deployment is configured.
Setup effort
One command plus credentials — claude mcp add vertica --scope user -- npx -y @hechtcarmel/vertica-mcp@latest --env-file <YOUR_ENV_FILE>, then supply credentials
USE CASES
01
Getting the shape of an unfamiliar schema before writing a query
Column types and constraints decide whether a query is even valid, and view definitions often answer the question outright.
02
Pulling a result set too large to come back in one response
Batching is the difference between an answer and a truncated payload; the batch size and the row ceiling are both yours to set.
03
Working out why a query is slower than the data volume suggests
In Vertica the physical layout is the projection, and it is what a slow scan usually traces back to.
6 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
get_table_structure
Get detailed structure information for a specific table including columns, data types, constraints, and metadata.
list_indexes
List indexes (projections) for a specific table in Vertica with column and uniqueness information
list_tables
List all tables in a schema with metadata
list_views
List all views in a schema with their definitions and metadata
stream_query
Stream large query results in batches to handle datasets efficiently. Query restrictions depend on configuration.
BEFORE YOU INSTALL
What to know first
Tools you get
6 tools when we connected. VERTICA_READONLY_MODE defaults to true, limiting execute_query to SELECT, SHOW, DESCRIBE, EXPLAIN and WITH. Use stream_query above about 10,000 rows.
Requires
- npx on PATH
- Node.js >= 18.0.0
- VERTICA_HOST, VERTICA_DATABASE and VERTICA_USER in the .env file (VERTICA_PORT defaults to 5433; VERTICA_PASSWORD is optional)
Where it runs
All 3 documented clients run the server on your machine.
CONTEXT COST
Light Standing cost
Rated by the size of the 6 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
claude mcp add vertica --scope user -- npx -y @hechtcarmel/vertica-mcp@latest --env-file <YOUR_ENV_FILE>MCP Clients
Claude Code
Create your .env file with connection details: VERTICA_HOST, VERTICA_PORT (5433 by default), VERTICA_DATABASE, VERTICA_USER and VERTICA_PASSWORD.
Then run the command below, with --env-file pointing at that file.
CLAUDE CODE · TERMINAL
claude mcp add vertica --scope user -- npx -y @hechtcarmel/vertica-mcp@latest --env-file <YOUR_ENV_FILE>Cursor
CURSOR · JSON
{
"mcpServers": {
"vertica-mcp": {
"command": "npx",
"args": [
"@hechtcarmel/vertica-mcp",
"--env-file",
"/Users/<YOUR_USERNAME>/.cursor/vertica.env"
]
}
}
}Create ~/.cursor/vertica.env with the connection details — VERTICA_HOST, VERTICA_PORT (5433 by default), VERTICA_DATABASE, VERTICA_USER and VERTICA_PASSWORD — and point --env-file at it.
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"vertica-mcp": {
"command": "npx",
"args": [
"@hechtcarmel/vertica-mcp",
"--env-file",
"<YOUR_ENV_FILE>"
]
}
}
}--env-file points at a .env file holding the connection details: VERTICA_HOST, VERTICA_PORT (5433 by default), VERTICA_DATABASE, VERTICA_USER and VERTICA_PASSWORD.
