Skip to main content

Flow Steward Documentation

Operate AI workflows with clear human control.

Flow Steward runs, reviews, and coordinates AI-assisted workflows while keeping approvals, jobs, tools, skills, and extensions visible to the operator.

Start ReadingQuick Start

Documentation map

Start with the path that matches your role.

User Guide

Use the dashboard, projects, workflows, jobs, agents, datasets, MCP catalogs, extensions, and lifecycle controls.

Admin & Operations

Administer users, providers, security policy, audit evidence, events, queues, retries, and monitoring.

Deploy & Run

Install, update, observe, back up, recover, and troubleshoot production-like Flow Steward stacks.

Integrations & Extensions

Connect AI clients, external MCP catalogs, webhooks, data sources, provider callbacks, and extension bundles.

Developer Platform

Work on the app codebase, runtime, frontend, database, tests, MCP internals, and extension points.

API & Reference

Look up CLI commands, API surfaces, MCP tools, workflow step kinds, installer options, config, and upgrades.

First release shape

A public docs shell that can grow with the product.

This first cut uses docs-as-code: content lives as Markdown in the site repository, builds to static files, and can be deployed independently from the Flow Steward application.

  • Start Here routes builders, admins, operators, developers, extension authors, and MCP users to the right lane.
  • Admin and deploy runbooks separate production operations from everyday product usage.
  • Integration docs give MCP and extension work a discoverable home without implying unsupported embedding features.
  • Maintainer notes preserve audit coverage and documentation debt outside the primary customer path.