Operator Reference¶
Ralph Workflow is a free and open-source AI agent orchestration system built around a simple Ralph-loop core. That simple core composes into a stronger composable workflow system for substantial, well-specified repo work, and the default workflow is already strong enough to start with before you customize anything.
Ralph Workflow is the operating system for autonomous coding: a simple core loop that scales into a stronger workflow for substantial, well-specified repo work. This reference section is for day-to-day operation after that product story is already clear.
This section is for people running Ralph Workflow in real projects. Use it when you need commands, config, tool behavior, or web-access details rather than the product pitch.
Ralph Workflow keeps the core loop simple so the operating details stay teachable. It works best when you plug it into coding agents you already trust instead of rebuilding your setup or handing over provider secrets by default. If you need runtime internals or Python implementation details, use Developer Reference instead.
What lives here¶
CLI Reference — commands, flags, and sub-commands
Configuration Reference — config files, precedence, and common workflow knobs
Advanced Pipeline Configuration — workflow graph, phases, counters, recovery, fan-out
Advanced Artifact Configuration — artifact contracts, decision vocabularies, summaries
Advanced MCP Configuration — MCP servers, search, crawl, and web integrations
End-User Stories — common user goals and the shortest next doc for each one
MCP Tools — the built-in tool surface exposed to agents
Local Web Access — search, visit, and crawl behavior
Common operator questions¶
I need to edit
ralph-workflow.toml→ Configuration ReferenceI do not know whether to edit global config, local config, or pipeline policy → Configuration Reference
I need advanced docs for
pipeline.toml→ Advanced Pipeline ConfigurationI need advanced docs for
artifacts.toml→ Advanced Artifact ConfigurationI need advanced docs for
mcp.toml→ Advanced MCP ConfigurationI want the shortest docs path for my use case → End-User Stories