We build hosted memory for AI coding tools. One service, every tool, persistent context across sessions.
AI coding tools are powerful but forgetful. Every new session starts from scratch. Claude doesn't know what ChatGPT learned. Cursor doesn't remember what Claude Desktop was told. Your context, decisions, and preferences are lost every time a conversation ends.
MemPalace Cloud runs the open-source MemPalace memory engine as a hosted service. Your AI tools connect via the Model Context Protocol (MCP) and share a single memory layer — facts, decisions, project context, and lessons learned persist across sessions and across tools.
Two minutes from signup to persistent memory. No self-hosting, no ChromaDB, no embedding models to manage.
The MemPalace engine is MIT-licensed. If MemPalace Cloud disappears tomorrow, you can self-host the exact same stack. Your data is exportable at any time. We earn your trust by making it easy to leave.
MemPalace Cloud is a community-hosted instance of the open-source MemPalace engine, which is MIT-licensed. We are not officially affiliated with the upstream project or its maintainers — we operate transparently within the license terms.
We contribute back. Current upstream work:
All our contributions are visible on GitHub.
MemPalace Cloud is built by JC Software GmbH, a software company based in Hannover, Germany. Founded by Christian Schnatz — developer, architect, and the person who answers your support emails.
Questions, feedback, or partnership ideas: [email protected]