HTTP API (Scalar)
The /api/v2/ control-plane API, Airflow 3.2.x-compatible, as an interactive Scalar reference generated from the OpenAPI spec.
Reference
References for every Leoflow surface — the HTTP API, CLI, Go packages, Python runtime, configuration, and MCP server.
References for every Leoflow surface. The HTTP API, CLI, Go, and Python references
are generated from source on every push, so they never drift from the code. The
Configuration page is hand-maintained against
internal/config — treat the server source as the final authority.
HTTP API (Scalar)
The
/api/v2/ control-plane API — Airflow 3.2.x-compatible — as an interactive Scalar reference.
Open the API reference →
CLI reference
Every leoflow command and flag, generated from Cobra.
CLI commands →
Go packages (GoDocs)
GoDocs for the control plane, scheduler, executor, agent, and storage — one page per package, each symbol linking to source.
Go packages →
Python runtime API
The leoflow_runtime helpers your DAG code imports (XCom, staging paths) — plus the full generated docstring reference.
Python runtime →
Configuration
The LEOFLOW_* environment variables and config keys for the server.
Configuration →
MCP server
The Model Context Protocol server — read-only Tools and Resources that expose Leoflow to AI agents.
MCP server →The Glossary collects the core vocabulary (DAG, DagRun, XCom, Executor, …).
CLI reference
Every leoflow command and flag, generated from Cobra.
Go packages (GoDocs)
GoDocs for the control plane, scheduler, executor, agent, and storage — one page per package.
Python runtime API
The leoflow_runtime package that runs your task callable inside the container and bridges its return value to XCom.
Configuration
The LEOFLOW_* environment variables and config keys for the server.
Glossary
The core Leoflow vocabulary — DAG, Task, TaskInstance, DagRun, XCom, Executor, Agent — kept compatible with Airflow.
MCP server
The Leoflow MCP server now has its own top-level section.