Conversational PostgreSQL Agent with Visuals, Multi-KPI & Data Editing

This n8n workflow enables natural language interaction with PostgreSQL databases, allowing users to ask complex questions and receive comprehensive responses including summaries, QuickChart visuals (bar/pie charts), and multi-KPI analysis in a single prompt. It supports data editing through secure inserts and updates, all orchestrated by an intelligent agent using n8n's Model Context Protocol (MCP). The agent plans multi-step reasoning (up to 30 steps) to handle diverse tasks efficiently.

Key benefits include time-saving automation for finance teams, eliminating manual SQL writing, reducing errors with validation, and providing dashboard-ready Markdown outputs. It integrates Claude 3.5 Haiku for planning and tool-calling, DeepSeek for SQL/QuickChart generation, and modular components like checkdatabase subflow for safety. Estimated cost per run is just ~$0.02, making it scalable for production use.

Ideal use cases: Financial analysts querying sales KPIs with charts, marketers updating campaign data via chat, or operations teams generating reports and edits without coding. Models are swappable (e.g., OpenAI, Gemini), ensuring flexibility. Security features like parameterized queries and no raw SQL prevent injections, making it enterprise-ready.