IceShore

Make your AI better at system maintenance. IceShore is a configuration-management platform for teams whose systems have multiplied as AI sped up development. Alongside each system's architecture, it prompts your AI to organize what the system does for the business (its use cases) and serves the result back to it over MCP. Your AI can then explain a change's impact on business operations, on processing, and on external systems itself, instead of guessing from the code alone. The auto-generated architecture diagrams also let you check how well the AI understands each system. For example, ask your AI "If we stopped the medical-records database, who would be affected?" It reads the use cases behind the architecture and answers with the people and business processes that depend on that database. No account needed to try it: list_samples, get_content and get_guide work anonymously on the public sample workspaces. Sign in (free for a single person) to work with your own systems. Endpoint: https://mcp.iceshore.ai/mcp (Streamable HTTP). Auth: OAuth 2.1.

Documentation

How to connect your AI to IceShore

Connect IceShore to your AI. Once connected, you can create and manage your system's architecture (ADL — Architecture Description Language, which describes business meaning) simply by instructing your AI. Don't have an AI coding tool yet? See the official guides here.

Connect

This is the name your AI will use to refer to this connection. If you connect separately per system or per client, pick a name that's easy to tell apart.

To check the connection, ask your AI “check whether iceshore is available.” If it recognizes the tool, you're connected.