snowChat
Basic Information
snowChat is an application that lets users query Snowflake data using natural language instead of writing SQL. The project provides a Streamlit user interface where users type questions and the system generates and executes SQL against a configured Snowflake database. It integrates large language models to translate conversational input into precise SQL, retains conversational memory for context-aware follow ups, and uses an agent-based architecture to manage interactions and tool usage. The repo includes tooling to ingest schema documentation into embeddings, store indexes, and cache Snowflake responses using Cloudflare KV and Supabase components. Installation and runtime rely on environment secrets provided in a secrets.toml file and running ingest.py to build embedding indexes followed by starting the Streamlit app.
Links
App Details
-
$$$$ -
$$$$