Docs
Start from the kind of collection you need: guides, concepts, techniques, or documentation. This page points you to the right part of the reference and gives a few strong starting pages.
Quick routes
Use these entry points when you know how you want to explore, but not the exact page yet.
Guides
Guide pages walk through practical CLI workflows and how to apply the reference in day-to-day work.
- Cursor Dynamic Workflows
Use Cursor as an MCP host for you-agent-factory Factory Session and dynamic-workflow tools.
- Getting Started
Install you-agent-factory, start a factory, and submit your first work.
- Using you-agent-factory for Loops
Run iterative factory loops that keep work cycling until a goal is done.
- Write-Review Loops
Use you-agent-factory for write → review → revise loops with a bounded retry path.
Concepts
Concept pages explain broader ideas that span many factory workflows, such as harnesses, loops, and bottlenecks.
- Bottlenecks
A factory bottleneck is the stage or resource that caps overall you-agent-factory throughput even when other stages have spare capacity.
- Checklist
A factory planner checklist is a live outcomes and workstream board that shows done, active, ready-next, and intentionally held work.
- Compaction
Compaction shrinks retained context—history, transcripts, or stream windows—so long-running agent or factory work can continue within limits.
- Harness
A harness is the agent runtime the factory drives, such as a coding agent under a shared factory process.
- Loop
A factory loop keeps iterating under you-agent-factory until the goal is done.
- MCP
Model Context Protocol: the host↔server protocol that exposes named tools, including Factory Session tools via you mcp serve.
Techniques
Technique pages explain named methods and practices you can apply across CLI workflows and documentation tasks.
- Classify-Execute
Classify an item into a known class, then run the specialist execute path for that class.
- Fusion
Fusion is a you-agent-factory technique that runs two model passes on the same request: a first pass drafts an answer and a second pass refines it.
- Planner-Executor
Planner-executor separates planning work into reviewable slices from executing those slices so independent factory streams can move in parallel.
- Planner-Executor In Action
Practice planner-executor as a multi-model cost tradeoff: when a larger planner and smaller executor win over one shared model.
- Ralph
Ralph is an autonomous one-story-per-iteration execution loop driven by a product requirements backlog.
- Worker-Adviser
Worker-adviser separates an adviser role that recommends or critiques from a worker role that performs the work.
Documentation
Documentation pages cover reference material for the CLI surface, including contracts, conventions, and supporting detail.
- Agent worker
Configure an AGENT_WORKER for agent-loop execution: model and provider identity, agentTools policy, permissions, resources, and AGENT_RUN workstation companionship.
- Agent-run workstation
Configure type = AGENT_RUN for agent worker dispatch and OpenCode agent overrides through an AGENT_WORKER backend.
- API
Reach and use the you-agent-factory local HTTP API: default base URL, factory-running precondition, session-oriented flow, and the path to the operations catalog.
- Architecture of system
How you-agent-factory is structured: factory topology, work, workstations, workers, resources, and live FactorySession runtime.
- Classifier workstation
Configure type = CLASSIFIER_WORKSTATION for classification routes and classifier-specific output restrictions through a HOSTED_WORKER companion.
- CLI
Install the you-agent-factory CLI and look up the core you commands for starting a factory, submitting work, checking sessions, and reading packaged docs.