QueryPilot runs on your own machine — the client starts it, so there is no endpoint to ping. 39 installs a week from pypi. Last commit 12 Jul 2026.
Safe SQL gateway for AI agents: SELECT-only validation, access policies, masking, audit trail, evals
Today is the operative word: we check QueryPilot every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add querypilot -- uvx querypilot
{
"mcpServers": {
"querypilot": {
"args": [
"querypilot"
],
"command": "uvx"
}
}
}
[mcp_servers.querypilot]
command = "uvx"
args = ["querypilot"]
{
"mcpServers": {
"querypilot": {
"args": [
"querypilot"
],
"command": "uvx"
}
}
}
{
"mcpServers": {
"querypilot": {
"args": [
"querypilot"
],
"command": "uvx"
}
}
}
This one needs environment variables set before it will start:
QUERYPILOT_DATABASE_URL (SQLAlchemy-style database connection URL, e.g. sqlite:///demo.db or postgresql://user:pass@host:5432/db. Equivalent to --database-url.), QUERYPILOT_DIALECT (SQL dialect for the connector. Equivalent to --dialect.), QUERYPILOT_MAX_ROWS (Maximum rows returned per query. Equivalent to --max-rows.), QUERYPILOT_TIMEOUT_SECONDS (Query execution timeout in seconds. Equivalent to --timeout-seconds.), QUERYPILOT_ACCESS_POLICY_JSON (JSON-encoded AccessPolicy (blocked_columns, allowed_columns, row_filters, masking_rules) enforced before execution. Equivalent to --access-policy-json.), QUERYPILOT_MCP_TRANSPORT (MCP transport used by `querypilot mcp`. Equivalent to --transport. Most MCP clients should leave this at stdio; streamable-http is for clients that speak Streamable HTTP directly.).
The author declared them in the registry entry; get the values from the project itself.
Safe, read-only Postgres and MySQL access for AI agents. Audit log + column-level controls.
Safe, read-only database access for AI agents. 5 databases, batch queries, SQLGlot.
Fail-closed, read-only PostgreSQL and MongoDB access for AI agents over MCP.
Safe-by-default SQL guardrails for AI agents: AST-checked queries, per-table policy, audit log.
Safe, audited PostgreSQL operations for AI agents: queries, migrations, EXPLAIN, containers
Project tracker for AI agents: roadmap, periods, tasks and a validation gate, in one SQLite file
Let coding agents query any SQL database safely. Read-only by default and scoped by your policies.
Safety-railed database access for agents: Postgres, MySQL, Redis. Read-only by default.
Answers built from our own checks of this server.