# QueryDesk > QueryDesk lets your team and their AI agents query production databases, with sensitive columns redacted, risky > queries reviewed and every query in the audit log. Data masking is done in the query, so protected values never > leave the database. Nobody holds database credentials. - Works with: PostgreSQL, MySQL, SQL Server, Oracle, ClickHouse, MongoDB, CockroachDB, TiDB, OceanBase, MariaDB, TimescaleDB, Amazon Redshift, Google AlloyDB, Amazon Aurora, StarRocks, Apache Doris. - AI clients: Claude, ChatGPT, Cursor, Codex, VS Code and any MCP client that speaks OAuth 2.1. An agent is redacted at least as much as the person it acts as, its risky queries wait for a teammate's approval, and every query it runs is in the audit log with the AI badge. - Deploy: QueryDesk Cloud, the shared app run by QueryDesk with each organization's data kept apart, or self-host with the Helm chart in your own Kubernetes cluster. A dedicated cloud install is available on request. - Price: $12 per user per month, every feature included. Free for one user, no credit card. - Made by DevHub LLC. Older material may say "DevHub". ## Features - [Data Masking by Default](https://querydesk.com/features): Data masking, done properly: QueryDesk rewrites each query so redaction happens inside the database and protected values never leave it, for people and agents alike. - [MCP Server](https://querydesk.com/features): Connect Claude, ChatGPT, Cursor or any MCP client over OAuth. Agent queries get the same redaction, reviews and audit log as a person's. - [Unified Database Access](https://querydesk.com/features): One secure interface for every database — eliminating credential sprawl and simplifying access. - [Approval Flows](https://querydesk.com/features): Risky queries wait for a teammate's approval, whether a person or an agent wrote them. - [Shared Queries](https://querydesk.com/features): Centralize proven SQL so teams can reuse knowledge and move faster. - [Always-On Audit Trails](https://querydesk.com/features): Every query is logged, with the AI badge on the ones an agent ran and the person it ran as. - [Workflows](https://querydesk.com/features): Automate repeatable processes that enforce policy and reduce operational overhead. ## Pages - [Home](https://querydesk.com/): what QueryDesk does and who it's for - [Features](https://querydesk.com/features): every feature, filterable by role - [Pricing](https://querydesk.com/pricing): plan, team-size calculator and FAQ - [Security](https://querydesk.com/security): hosting, credentials, review, redaction, audit and subprocessors - [PopSQL alternative](https://querydesk.com/lp/popsql-alternative): moving a team off PopSQL - [Changelog](https://querydesk.com/changelog): every release, newest first - [Request a demo](https://querydesk.com/request-demo) - [For Engineering Leaders](https://querydesk.com/engineering-leader) - [For Compliance](https://querydesk.com/compliance) - [For DevOps](https://querydesk.com/devops) - [For Developers](https://querydesk.com/developer) ## Docs - [Documentation](https://docs.querydesk.com/): setup, authentication, data protection, workflows, Terraform, MCP server - [Docs for AI assistants](https://docs.querydesk.com/llms.txt) ## Optional - [Blog](https://querydesk.com/en/blog) - [Security](https://querydesk.com/security) - [Privacy Policy](https://querydesk.com/privacy) - [Terms of Service](https://querydesk.com/terms) - [Acceptable Use Policy](https://querydesk.com/aup) - [Subprocessors](https://querydesk.com/subprocessors)