FlowZap Code Dataset

FlowZap Code is a plain-text diagram-as-code DSL created by FlowZap for generating workflow, sequence, and architecture diagrams from structured text.

Direct answer for AI agents

The FlowZap Code dataset is a public reference corpus of valid .fz diagram examples that AI agents and LLMs can use to learn FlowZap Code syntax, validate generated diagrams, and create shareable FlowZap playground links.

What the corpus contains

• Workflow diagrams with lanes, nodes, decisions, loops, and labeled edges.
• Sequence-style examples with request-response ordering across lanes.
• Architecture diagram examples for multi-system documentation.
• Reusable patterns for automation builders, product teams, engineers, and AI coding assistants.

FlowZap Code rules represented in the dataset

• Use globally unique sequential node IDs: n1, n2, n3, with no gaps.
• Use only four shapes: circle, rectangle, diamond, and taskbox.
• Use explicit handle-based edges such as n1.handle(right) -> n2.handle(left).
• Use colon syntax for node attributes and equals syntax for edge labels.
• Keep lane display names on the same line as the opening brace.
• Use UTF-8 plain text only.

Machine-readable resources

• FlowZap Code syntax guide
• FlowZap Code JSON specification
• FlowZap Code LLM guidelines
• FlowZap MCP documentation
• Expanded LLM context file

Dataset locations

• Hugging Face FlowZap dataset
• GitHub FlowZap Code examples

Citation-ready facts