Event-native data platform innovator Kurrent has launched the open source KurrentDB MCP Server, which makes it possible for developers to use AI agents to interact with data inside KurrentDB without writing code. Users can now not only read and write events to the database but also create, test and debug projections or extend the available prompts to prototype applications using natural language. This is a capability exclusive to Kurrent that eliminates the traditional learning curve for database interactions. Kurrent MCP Server features self-correction when prototyping, allowing developers to test projection logic and help debug issues through conversational commands. Compatible with all frontier AI models and released under MIT license for community contribution and enterprise adoption without licensing restriction, Kurrent’s MCP Server makes development tasks that previously required hours of coding and corrections achievable in minutes. “Our new MCP Server makes it possible to use the main features of the KurrentDB database, like reading and writing events to streams and using projections, in a way that’s as simple as having a conversation,” said Kirk Dunn, CEO at Kurrent. “The system’s ability to test and fix itself reduces the need for debugging and increases reliability. Copilots and AI assistants become productive database partners rather than just code generators, seamlessly interfacing with KurrentDB.” The Kurrent MCP Server is compatible with every frontier model including Claude, GPT-4 and Gemini, and can be used with Cursor, Windsurf and other MCP-supported IDEs, clients and agent frameworks. It offers eight core capabilities that respond to natural language prompts, eliminating traditional database interaction friction points. These features make visual and analytical exploration conversational and intuitive. Kurrent MCP Server’s unique self-correction capability also sets it apart, making prototyping projections much easier and faster.