-
security-
license-
qualityAn MCP server that enables natural language querying of Supabase PostgreSQL databases using Claude 3.7, allowing users to inspect schemas, execute SQL, manage migrations, and convert natural language to SQL queries.
Utilized for visualizing and documenting the prompt-to-SQL workflow in the application
Integrates with Ollama to translate natural language queries into SQL, enabling non-technical users to query databases without SQL knowledge
Enables querying PostgreSQL databases with both raw SQL and natural language queries, with results available for analysis, charting, and export
Powers the user interface with interactive dashboards for visualizing query results, generating charts, and exporting data to CSV
🚀 MCP-Powered Streamlit Dashboard with Ollama + PostgreSQL
A modern data analysis tool that lets you:
Query your PostgreSQL database with natural language
Run SQL manually or generate it using LLMs (Ollama)
Analyze data with charts, statistics, and export to CSV
🧠 Prompt-to-SQL Flow
graph TD
A[🧑 User types a data question] --> B[Streamlit sends prompt to Ollama API]
B --> C[Ollama generates SQL query as text]
C --> D[Streamlit extracts the SQL]
D --> E[Streamlit sends SQL to MCP server]
E --> F[MCP executes query on PostgreSQL]
F --> G[Results returned to Streamlit]
G --> H[📊 Results shown + Chart + CSV Export]
📦 Architecture
Streamlit – UI + charting
MCP (FastMCP) – Tools/resources for SQL query and table listing
PostgreSQL – Stores your company data
Ollama – LLM that translates natural language prompts to SQL
📌 Features
✅ Natural language → SQL
✅ Charting (bar/line/time series)
✅ CSV download
✅ Statistical summary
✅ Prompt explainability with raw output
✅ Auto-detect date/time fields
✅ LLM integration with llama3 (configurable)
🛠️ Getting Started
git clone <this-repo>
cd postgres-mcp-server
docker-compose up --build
Access UI: http://localhost:8501
MCP API: http://localhost:3333/mcp
⚙️ Env Configuration
MCP_API_URL=http://mcp-server:3333/mcp
OLLAMA_URL=http://ollama:11434/api/generate
📤 Prompt Example
“List departments with average salary > 50000”
👉 Translated to SQL:
SELECT department, AVG(salary) FROM employees GROUP BY department HAVING AVG(salary) > 50000;
Why is this a good use case for MCP?
🔗 MCP makes it dead simple to expose structured tools like SQL queries to LLMs. 🎯 Agents can discover and call your tools without hardcoding logic. 💬 You get the best of both worlds — interpretability, flexibility, and control.
Whether you're building internal tools, research dashboards, or intelligent agents — this pattern is reusable, secure, and 100% local.
📄 License
MIT
This server cannot be installed
Related Resources
Related MCP Servers
- -securityFlicense-qualityA FastMCP server that provides natural language interaction with MS SQL databases, enabling users to query data, list tables, describe structures, and execute database operations through a conversational AI interface.Last updated -
- -securityAlicense-qualityMCP-Server from your Database optimized for LLMs and AI-Agents. Supports PostgreSQL, MySQL, ClickHouse, Snowflake, MSSQL, BigQuery, Oracle Database, SQLite, ElasticSearch, DuckDBLast updated -503Apache 2.0
- -securityAlicense-qualityA FastMCP-based server that enables executing Microsoft SQL Server data queries and table structure queries through a standardized MCP protocol interface.Last updated -5MIT License