Model, simulate, query your system — together, with AI.

Define your own metamodel — node types, relationship types, allowed connections — visually on a canvas. The metamodel is the scaffolding that keeps every instance valid and gives stakeholders a guided contribution surface they can't break.

Build models together as a team. Changes propagate instantly via WebSocket — every collaborator sees the same canvas in real time, with presence indicators showing who's working where.

Describe what you want in plain language and the AI assistant generates nodes, relationships, and structure that respect your metamodel constraints. It coaches you through modeling rather than replacing your judgment.

Every Metapad model is a typed knowledge graph that AI tools can query reliably — no hallucination, just structured facts. Export to standard formats, query via REST API, or connect any AI tool via MCP.

Turn your model into an executable specification. The M0 simulation layer runs scenarios over time so you can test policy changes, demand shocks, or org redesigns before they hit production.

Models are multilingual at the data level — labels, descriptions, and constraints can be authored in multiple languages and switched on the fly. Built for global teams from day one, not bolted on.

Keep sensitive models entirely in your browser — no server, no upload. Ideal for early-stage prototypes, regulated data, or air-gapped review with stakeholders before sharing more broadly.

Attach documents, images, and files to model elements. Embed them in markdown descriptions to weave structure and narrative together — the literate-models pattern in practice.

Send any stakeholder a link to an interactive read-only view of your model. They explore the canvas, drill into nodes, and form their own questions — no account required, no edit risk, no separate review tool.