An agent writing Ilograph without this reaches for remembered syntax and iterates until the file opens. Here it can read the authoritative property table first and then get told exactly which line failed and what to change. The project is unofficial and reads public Ilograph docs — check generated diagrams against the vendor's own documentation before you rely on them.
A documentation-and-validation companion for Ilograph diagrams. It fetches the official specification and documentation sections from Ilograph's own sources, checks a diagram in two stages — YAML syntax, then the Ilograph schema — and searches the live icon catalogue. It is a community project, not affiliated with Ilograph LLC.
- Documentation fetched by section — resources, relation and sequence perspectives, contexts, imports, icons, walkthroughs, the full tutorial and more — `fetch_documentation_tool`, `list_documentation_sections`
- The official specification with every property, its type and whether it is required — `fetch_spec_tool`
- Two-stage validation returning errors, warnings and info with locations and suggested fixes, plus a readable overall assessment — `validate_diagram_tool`, `get_validation_help`
- Worked example diagrams by complexity, each with its learning objectives and the patterns it demonstrates — `list_examples`, `fetch_example`
- Semantic search across the live icon catalogue, filtered by provider such as AWS, Azure or GCP — `search_icons_tool`, `list_icon_providers_tool`
- Health checks on the documentation and specification services with cache statistics — `check_documentation_health`, `check_spec_health`
Docker installed and running; the client launches `ghcr.io/quincymillerdev/ilograph-mcp-server:latest` with `docker run -i --rm`. No account and no key. Python 3.11+ if you build it yourself.
One command — docker run -i --rm ghcr.io/quincymillerdev/ilograph-mcp-server:latest