snowflake-mcp-lambda

What is snowflake-mcp-lambda?

A web application that enables non-technical users to upload CSV files and query them using natural language via a chat interface. It integrates with Google Gemini for natural language to SQL conversion and optionally uses Google OAuth for authentication.

How to use snowflake-mcp-lambda?

Clone the repository, run make setup to copy the environment file, edit .env with your Gemini API key, then run make dev-setup to start containers. Access the frontend at http://localhost:3000 and the backend API at http://localhost:8000. Use make test to run backend tests.

Key features of snowflake-mcp-lambda

• CSV upload with drag-and-drop and progress indication
• Schema preview with auto-detected columns and data types
• Natural language chat interface for querying data
• SQL generation via Google Gemini LLM
• Read-only queries (SELECT only) for security
• No data persistence; files deleted after session ends
• User-provided Gemini API key, never stored server-side
• Optional Google OAuth authentication

Use cases of snowflake-mcp-lambda

• Non-technical users analyzing CSV data without SQL knowledge
• Quick data exploration and ad-hoc queries via chat
• Prototyping data analysis workflows without database setup
• Secure data analysis with temporary file handling

FAQ from snowflake-mcp-lambda

What LLM does the application use?

It uses Google Gemini API for converting natural language questions to SQLite queries.

Are uploaded files stored permanently?

No, files are deleted after the session ends. The application does not persist data.

What authentication methods are supported?

Optional Google OAuth authentication is supported. Without it, users can still use the application.

What database is used for querying?

Uploaded CSV files are converted to in-memory SQLite databases for fast query execution.

What file formats are supported?

Currently only CSV files are supported. Future plans include Excel, JSON, and Parquet.