Choreo

Domain-specific diagram builders and graph analyzers on top of Yog.

Choreo is a family of Elixir libraries that let you model, analyze, and render complex systems as graphs. Instead of drawing boxes and arrows by hand, you write code. Instead of static pictures, you get live analysis — reachability, cycles, bottlenecks, threat generation, and more.

alias Choreo.Dataflow

# A dataflow pipeline with one line of analysis
pipeline =
  Dataflow.new()
  |> Dataflow.add_source(:sensor, label: "IoT Sensor")
  |> Dataflow.add_transform(:parse, label: "JSON Parser")
  |> Dataflow.add_sink(:db, label: "TimescaleDB")
  |> Dataflow.connect(:sensor, :parse, data_type: "raw bytes")
  |> Dataflow.connect(:parse, :db, data_type: "event")

Dataflow.Analysis.cyclic?(pipeline)      #=> false
Dataflow.to_mermaid(pipeline)            #=> Mermaid diagram

graph TD
  classDef default color:white
  parse[["JSON Parser"]]
  db["TimescaleDB"]
  sensor(["IoT Sensor"])
  style parse fill:#3b82f6
  style db fill:#f43f5e
  style sensor fill:#10b981
  sensor -->|raw bytes| parse
  parse -->|event| db

Installation

Add choreo to your mix.exs:

def deps do
  [
    {:choreo, "~> 0.14"}
  ]
end

Detailed Guides

Choreo supports 16 different artifact modeling vocabularies. They are categorized and detailed in the following guides and references:

1. Architecture & Design Modeling

   • System Architecture (Choreo)
   • Cloud Network Topology (Choreo.Infrastructure)
   • C4 Model Architecture (Choreo.C4)
   • STRIDE Threat Modeling (Choreo.ThreatModel)
   • Domain-Driven Design & Event Storming (Choreo.Domain)
   • Database ERD Design (Choreo.ERD)
   • UML Class & Struct Diagrams (Choreo.UML)
   • Requirements Engineering & Traceability (Choreo.Requirement)

2. Behavior & Flow Modeling

   • Finite State Machines (Choreo.FSM)
   • Sequence Diagrams (Choreo.Sequence)
   • Saga Task Orchestration (Choreo.Workflow)
   • Dataflow Pipelines (Choreo.Dataflow)

3. Data & Structure Modeling

   • Software Dependency Graphs (Choreo.Dependency)
   • Decision Trees (Choreo.DecisionTree)
   • Concept Mapping / Mind Maps (Choreo.MindMap)
   • Project Task Planning (Choreo.Planner)

4. Analysis Algorithms Reference

   • Catalog of graph algorithms used by each analysis module
   • Function-to-algorithm mapping for every diagram type

5. Lab DSL & Sketch Syntax (Cheatsheet)

   • Experimental Livebook-friendly macro DSLs for sketching models across all domains
   • Compact syntax with variable binding, directional operators (~>), and pipe modifiers

Interactive Notebooks

The livebooks/ directory contains 30+ interactive walkthroughs, integration examples, and complete system designs that you can run directly in Livebook:

• livebooks/guides/ — step-by-step introductions to each Choreo module and diagram type:
  • Walkthroughs for all 15 domains: C4, Dataflow, Sequence, Threat Model, ERD, FSM, Workflow, Dependency, Decision Tree, Domain Modeling, Mind Map, Planner, Requirement, Infrastructure, and UML.
  • System Design Walkthrough — end-to-end design session composing C4, Dataflow, Workflow, and Threat Model.
  • Advanced Analysis Walkthrough — centralities, heatmaps, cycle detection, and graph decomposition.
  • Lab Visualization Walkthrough — interactive Siren pan/zoom controls and Sketch (Excalidraw) whiteboards.
• livebooks/projects/ — complete, multi-perspective system design notebooks:
  • API Gateway System Design — C4 context/container, dataflow, token-refresh saga workflow, STRIDE threat model, and review prompts.
  • Real-Time Vehicle Tracking System — IoT ingestion platform with ERD, C4, dataflow, sequence, trip saga, vehicle FSM, threat model, and requirements.
  • Distributed Web Crawler — web crawler architecture with C4, crawl dataflow, fetch saga, ERD, threat model, and requirements traceability.
• livebooks/integrations/ — notebooks that bridge Choreo with third-party tools, formats, and data sources:
  • Finitomata — design FSMs in Choreo, run them with Finitomata, and analyze them back in Choreo.
  • GitHub Issues — turn a public repo's issues into a Planner and Mind Map.
  • Hex Dependencies — crawl any Hex package's dependency tree.
  • Mix Xref — visualize and analyze internal Elixir project dependencies.
  • Ecto Schema ERD — introspect Ecto schemas and render them as an interactive ERD.
  • Requirements Exchange — import and analyze requirements from CSV, JIRA, and IBM DOORS modules.
  • YogEx Algorithm Selector — interactive decision tree guiding you to the right YogEx graph algorithm.
• livebooks/extending_choreo/ — tutorials that walk through adding new diagram types, analysis vocabularies, and protocol implementations to Choreo:
  • Git Graph — extend Choreo with a Mermaid gitGraph renderer, from builder to real git log adapter.
  • Call Graph Analysis — extend Choreo with an analysis-first call graph: dead functions, cycles, hotspots, and impact analysis.
  • FSM Viewable — add Choreo.Viewable support to Choreo.FSM so it can be zoomed, focused, and filtered.

Graph Analysis & Heatmaps

Choreo provides graph analysis tools to identify "hotspots" in your architecture, workflows, and pipelines. Use heatmap/2 to automatically color nodes based on importance or performance metrics.

Themes & Rendering

All modules render to DOT (Graphviz) and Mermaid.js via a shared theming pipeline.

# DOT output (Graphviz)
Choreo.to_dot(system, theme: :default)
Choreo.to_dot(system, theme: :dark)

# Mermaid.js output (GitHub, GitLab, Notion, Livebook)
Choreo.to_mermaid(system, theme: :default)
Choreo.to_mermaid(system, theme: :ocean)

# Custom theme
theme = Choreo.Theme.custom(
  colors: %{database: "#ff0000", service: "#00ff00"},
  graph_bgcolor: "#0f172a",
  node_fontcolor: "white"
)
Choreo.to_dot(system, theme: theme)
Choreo.to_mermaid(system, theme: theme)

Interactive Livebook Widgets (Experimental)

Choreo provides two custom Livebook widgets for rich, interactive diagram visualizations. They are compile-time optional and load automatically when kino is present:

1. Choreo.Lab.Siren: An enhanced Mermaid.js renderer (Mermaid v11.x) with hardware-accelerated pan/zoom controls, floating zoom toolbar, dynamic fit-to-screen scaling, and automatic dark/light theme detection.

   mermaid_code = Choreo.to_mermaid(system)
   Choreo.Lab.Siren.new(mermaid_code, height: "600px")
   

2. Choreo.Lab.Sketch: Renders any Mermaid flowchart inside an interactive Excalidraw whiteboard. The diagram is converted to sketch elements on the fly, allowing you to double-click, draw notes, and scribble ideas directly on top of your Choreo models.

   Choreo.Lab.Sketch.new(mermaid_code, height: "600px")
   

MCP Server for Agents

Choreo ships with a zero-dependency stdio MCP server so agents can read, write, and verify system-design notebooks using the same validation logic as mix choreo.test_livebooks. Run it from the Choreo repo:

mix choreo.mcp

It exposes four tools over JSON-RPC:

Configuring a client

For example, in Claude Desktop add this to claude_desktop_config.json (replace /path/to/choreo with your local clone):

{
  "mcpServers": {
    "choreo": {
      "command": "sh",
      "args": ["-c", "cd /path/to/choreo && exec mix choreo.mcp"]
    }
  }
}

Example prompts

• "Initialize a system-design notebook for a real-time URL shortener at livebooks/projects/url_shortener_system_design.livemd."
• "Read the C4 Context and C4 Container sections from livebooks/projects/api_gateway_system_design.livemd."
• "Add a Dataflow section to the URL-shortener notebook showing clients → API → cache → database."
• "Verify the URL-shortener notebook and report any runtime errors or design issues."

CLI Tools

Choreo includes Mix tasks for diagram rendering and notebook verification:

# Render .choreo.exs files directly to Mermaid (.mmd) or Graphviz DOT (.dot)
mix choreo.render diagrams/system.choreo.exs --to mermaid --out priv/diagrams/

# Headlessly evaluate and validate all Livebook notebooks
mix choreo.test_livebooks

Testing

mix test

All modules ship with comprehensive ExUnit test suites covering builders, analysis, rendering, and doctests.

License

MIT