$ loading_
Control agent workflows with stateful primitives and persisted execution facts.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "flow-mcp" yet — see the docs or source repo.
Use flow-mcp to manage a content-generation agent workflow: start the task, pause at the human review step, resume after approval, and rewind to the previous step if the direction is wrong. Return the control primitives to call at each stage.
A clear sequence of workflow control calls showing when to finish, pause, resume, rewind, or abort.
I am building a data collection agent. Explain how to use flow-mcp to write completed steps, pending items, and actionable facts into structured state files so execution can continue after interruptions.
Guidance on state persistence, including what information to record and how to resume execution.
Design an agent orchestration plan where flow-mcp aborts the current workflow after three consecutive external API failures, rewinds to the latest safe checkpoint, and resumes later when the service recovers.
An exception-handling workflow describing when to abort, how to rewind, and how to continue after recovery.
Add agentic tools with iterative reasoning and tool use to apps
Expose local Flowise chatflows as MCP tools for listing and execution.
Combine structured reasoning and execution so agents can think, act, and trace decisions.
Orchestrate AI agent workflows with dependencies, parallel execution, and failure policies.
Start, stop, and monitor background shell processes without blocking chat.
Build, debug, and manage software tasks with natural language across LLMs.