Problem solved

Software teams often struggle to keep architecture diagrams synchronized with implementation and to provide AI coding agents with accurate architectural context. Viaduct addresses this by offering a C4 modeling environment plus a hosted MCP server that feeds architecture models to AI development tools, enabling agents to work from current system context, container, component, and code diagrams.

Use cases

• Create C4 system context, container, component, and code diagrams
• Model user journeys and contracts alongside software architecture
• Provide architecture context to AI coding agents via MCP
• Enable live collaboration on software architecture models
• Generate change sets that coding agents can implement
• Document software architecture for team communication
• Use as an alternative to PlantUML C4 for architecture diagramming
• Feed LLM-based development tools with current architecture context

Audience

• Developers
• Product Managers
• Enterprise
• Startups

Pros

• Combines C4 architecture modeling with a hosted MCP server for AI coding agents
• Supports system context, container, component, and code diagrams
• Includes live collaboration and change sets for team workflows
• Free pricing according to the listed category
• Feeds architecture context directly to AI development tools

Cons

• Requires MCP-compatible AI tools for agent integration
• Public information on licensing, export formats, and release history is limited
• Being a hosted web service means internet access is required
• No GitHub repository or self-hosted option is listed on the homepage
• Feature limits and long-term roadmap are not detailed on the homepage