Inwardis
Self-hosted architecture model that agents read layer by layer, search, build and correct.
- Transport
- Not stated
- Package
- —
- Registry id
- com.inwardis/inwardis
No install snippet on purpose. A working MCP config is a command, its arguments and an environment block — the last two are where API keys live, so this catalogue never stores them and cannot publish them. Follow the link above for the authors' own instructions.
Inwardis keeps systems as one hierarchical model with context, containers, components, decisions, and data. It is self-hosted and runs with Docker, ensuring the model never leaves the server. The built-in MCP server lets an AI agent read the model layer by layer, search it, and build and correct it.
The server exposes 55 tools on a default install. The render_view tool returns the diagram as SVG with a summary of what was drawn. Every change appears in the activity feed, and the agent can leave notes while comments work both ways.
A 30-day free trial includes every feature, including the MCP server.
Summary drafted from the project's own website. Every sentence is backed by text on that page and was reviewed before publishing.