Sending production database credentials or raw data to a cloud-based SQL platform creates an unnecessary security risk. Most modern data tools force you to trade privacy for convenience, pushing sensitive information to opaque backends.
I built Dory to stop this. It is an AI-native data studio that runs locally, keeping your data inside your environment at all times. You connect via SSH tunnels and execute queries directly on your infrastructure, whether you are using PostgreSQL, ClickHouse, or other database engines.
Instead of relying on hallucinations or ungrounded models, Dory uses table intelligence and deep schema awareness to provide accurate insights. It acts as a professional SQL workbench that bridges the gap between raw editing and natural language analysis without the bloat of enterprise SaaS.
You can find the project at https://getdory.dev. It is licensed under Apache 2.0, allowing for self-hosted deployments using Docker or macOS bundles. If you are tired of cloud-only lock-in risks, have you tried moving your analysis workflow closer to the data itself?