InsightForge AI
Retrieving workspace models...
Retrieving workspace models...
InsightForge AI orchestrates multi-agent LangGraph pipelines to explore database schemas, index semantic embeddings, auto-optimize query performance, and self-heal SQL bugs with zero-trust execution.
Designed from the ground up for modern data engineers, analysts, and developers. Every feature operates smoothly across all your connected databases.
Translates plain English queries into precise SQL by injecting exact table definitions, column types, foreign keys, and pgvector semantic matches into the AI prompt.
Click "Optimize" on any query to receive deep performance diagnostics, covering index recommendations, and rewritten SQL with WHERE/LIMIT safeguards.
Encountered a syntax error or typo in a column name? Click "Auto Fix" and our dedicated debug node analyzes database errors and corrects the SQL editor automatically.
Navigate live metadata with collapsible tables, column type badges, primary key (PK) / foreign key (FK) highlights, and interactive ER relationship views.
Connect external PostgreSQL instances, MySQL servers, or local SQLite database files (`.db`) right from the top navigation dropdown with zero hassle.
Toggle "Teaching Mode" to get clear pedagogical explanations of SQL concepts, JOIN mechanics, subqueries, and aggregation rules alongside every result.
Built to protect production data while unleashing AI productivity. Every query undergoes strict pre-execution validation before touching your database.
All AI and user-submitted SQL is parsed by `is_safe_select_query`. INSERT, UPDATE, DELETE, DROP, and multi-statement attacks are automatically rejected at the API edge.
Stateful graphs route queries between Generation, Optimization, and Debugging nodes, ensuring high accuracy and context preservation.
Background worker (`discover_and_index_background`) indexes table catalogs and builds semantic vector embeddings in pgvector within milliseconds of connecting.
Log in to access your registered databases & history