Skip to content
dinobase

Cursor

Dinobase integrates with Cursor via MCP. Once connected, Cursor’s AI agent can query all your synced business data while you code.

Install

Section titled “Install”
Terminal window
curl -fsSL https://dinobase.ai/install.sh | bash -s -- cursor

Installs Dinobase via uv, runs dinobase init, and writes CLI usage instructions to AGENTS.md in your current directory. Then connect your data sources:

Terminal window
dinobase add stripe --api-key sk_test_...
dinobase add hubspot --api-key pat-...
dinobase sync

See Connecting Sources for the full list of 100+ supported sources, and Syncing & Scheduling for background sync options.

Cursor’s agent can then run shell commands like dinobase info, dinobase describe stripe.customers, and dinobase query "SELECT ..." to answer questions about your data.

To re-run the setup step: dinobase install cursor from your project root (safe to run multiple times — replaces the existing block).

How it works

Section titled “How it works”

The agent will:

  1. Run dinobase info to discover available data
  2. Run dinobase describe <table> to understand columns and types
  3. Run dinobase query "<sql>" to execute SQL (DuckDB dialect)
  4. For mutations (UPDATE/INSERT), query returns a preview — the agent runs dinobase confirm <id> to execute

Alternative: MCP server

Section titled “Alternative: MCP server”

If you prefer MCP, run dinobase mcp-config cursor to get the JSON config, then add it to .cursor/mcp.json in your project root.

Available CLI commands

Section titled “Available CLI commands”
CommandDescription
dinobase infoDatabase overview — sources, tables, row counts
dinobase describe <table>Table schema, column types, and sample data
dinobase query "<sql>"Execute SQL queries (DuckDB dialect)
dinobase refresh [source]Re-sync a data source to get fresh data
dinobase confirm <id>Execute a pending mutation
dinobase cancel <id>Cancel a pending mutation

Next steps

Section titled “Next steps”