Skip to main content

Getting started

Start accelerating your development​

Quick Start​

  1. Provide a Git token — Configure your personal access token for the Git provider to enable project creation and repository access.

  2. Create a new project — Select a project baseline that defines your modeling vocabulary and implementation recipes. Choose your Git provider and repository location.

  3. Choose profiles — Your selected baseline includes profiles (C4, domain driven design, implementation, etc.) that define the modeling vocabulary. You can enable and disable them as needed.

  4. Design and specify your project — Use the modeling canvas to create architecture diagrams, define domain models, and document decisions. Your design becomes the single source of truth.

  5. Generate your code with coding assistants — Coding assistants (Bob, Cursor, GitHub Copilot) read your design specifications and recipes to generate implementation.

Common workflows​

Architecture and Design in Workbench​

Create / evolve system architecture — Use C4 modeling, DDD, and Event Storming profiles to design your system architecture. Define boundaries, components, and relationships in a structured, versioned design graph that serves as the single source of truth.

Dependency analysis — Visualize and analyze dependencies between components, services, and domains. Understand impact of changes and maintain architectural coherence across your system.

Knowledge / decisions — Capture architectural decisions (ADRs), document rationale, and track assumptions. Link decisions directly to design elements, ensuring context is preserved and accessible to both humans and AI.

Collaborative design — Business stakeholders, architects, and engineers work together in a shared workspace. Changes are versioned in Git, reducing coordination overhead and eliminating the need to rediscover intent every sprint.

Explore with Spotlight — Quickly search and navigate your design graph. Find elements, relationships, and decisions across projects. Understand system context without navigating through multiple diagrams.

AI-Assisted Implementation​

Generate code — Coding assistants read your design specifications and recipes to generate implementation that aligns with architectural decisions from the start. The first prompt reflects the agreed design, reducing rework and accelerating delivery.

Align with standards and patterns — Recipes (skills, commands, and rules) guide AI agents to apply your organization's coding guidelines, technology definitions, and implementation patterns automatically. Consistency is built in, not bolted on.

Less prompt, more quality — With explicit design context and implementation guidance, AI agents generate accurate code without repeated explanations. Developers review and refine rather than translate and re-explain, moving from idea to code with efficiency.

Next Steps​