Add Connection
SQL execution endpoints — databases Claude queries directly.
Saved Connections
Free-form context always injected into Claude's memory — business rules, data quirks, definitions, conventions.
When a keyword appears in a query, its associated context is injected automatically. Use for metrics, column names, or SQL snippets.
Saved SQL queries stored as vector embeddings. Claude surfaces these when answering similar questions. Built automatically from Looker/Databricks ingestion or saved manually via dante_save_pattern.
Generate semantic embeddings from dashboard charts and data sources. These embeddings help the AI find relevant past queries and patterns.
Use AI to strip BI-platform noise from raw SQL, improving embedding quality. This runs as a post-processing step on existing embeddings.
Find and merge semantically similar embeddings to reduce redundancy and improve retrieval quality.
Status
A snapshot of what dante knows right now — your active database connection and accumulated knowledge.
Project Config
Settings stored in .dante/config.yaml for this project. The default connection tells Claude which database to query when you don't specify one.