Keep your systems in mind

Mynden is architecture documentation for software teams. Describe each system as a C4 model and Mynden draws the diagrams. Its decisions, risks, threats, reviews and tasks are kept on the same system.

Mynden showing the Checkout system: its context diagram with customers, support agents, Catalogue and Stripe, and its description and tech stack beside it.

C4 models and diagrams

Describe a system's actors, containers, external systems and the relationships between them. Mynden draws the diagrams from that model, so they always match it.

The design of the Checkout system in Mynden: its actors, its containers with their technology and open threats, and the external systems it depends on.

Architecture decisions

Record each decision as a numbered ADR with its context and consequences, next to the parts of the design it concerns. Superseding a decision links the old one and the new one.

A decision record in Mynden, ADR-001 Keep orders in PostgreSQL, with its context, decision, consequences and a review finding.

Risks and threats

Keep a risk register with a probability and impact matrix, and model STRIDE threats per part of the design. Each one has an owner, a mitigation and a status.

The risks of the Checkout system in Mynden: a probability and impact matrix and a list of three risks with their scores, statuses and owners.

Also in Mynden

Architecture reviews: Plan a review, record its findings by severity, and link each finding to the decision, risk or threat it is about.
Tasks: A board per system. Create a task straight from a decision, risk, threat or finding, with an owner and a due date. My tasks lists yours across workspaces.
Health: Per system: how complete the design is, open risks by score, open threats, the last review and overdue tasks. The workspace dashboard shows all systems side by side.
Landscape: Link an external system to a system in your workspace and the landscape map shows how all your systems depend on each other.
Activity: Every change is recorded with who made it and when, per system and per workspace.
Import and export: Import a Structurizr DSL or JSON file. Export a system as Structurizr DSL, as JSON, or as Markdown with the diagrams as SVG.
Search: Ctrl K finds any system, element, decision, risk, threat, review, task or chat you can read.
Workspaces and roles: One workspace per team or organisation. Members are owners, editors or viewers.

AI assistant and MCP

The assistant

Ask questions about one or more systems. It reads their design, decisions, risks, threats and reviews. It can draft a design from a conversation and suggest risks and threats. Changes it proposes come as a change set you review before applying.

MCP server

Connect Claude Code, Claude Desktop or another MCP client. It signs in as you and can read and change what your role allows in each workspace.

Who uses it

Architects and tech leads: Model the systems they own and keep decisions, risks and reviews current.
People new to a system: Read how it fits together and why it was built that way.
Engineering managers: See the landscape and which systems need attention.
AI agents: Read and update the same model over MCP, with the permissions of the person who connected them.

Start with one system

Create a workspace, register a system and describe its design, or import a Structurizr file you already have.