Is LoopCraft free to use?

Yes. LoopCraft offers a free tier for designing and exporting AI agent loop patterns. A Pro plan is available for advanced features.

What is Loop Engineering?

Loop Engineering is designing the control flow for AI agents: how they plan actions, observe results, and iterate. LoopCraft includes 11 pre-built patterns based on research from Anthropic, OpenAI, and others — including ReAct, Plan-Execute, Evaluator-Optimizer, Multi-Agent, and more. Design loops visually and export to LangGraph, OpenAI Agents SDK, AutoGen, or CrewAI.

Which frameworks can I export to?

LoopCraft exports to LangGraph (Python), OpenAI Agents SDK (TypeScript), AutoGen (Python), and CrewAI (Python). Each export produces working code you can run directly in your project.

Do I need to write code?

No. LoopCraft is a visual builder — you design loop patterns with a drag-and-drop interface and configure nodes without writing code. Code export to LangGraph, OpenAI SDK, AutoGen, and CrewAI is generated automatically and is optional.

What loop patterns are supported?

LoopCraft supports 11 agent loop patterns: ReAct (think-act-observe cycle), Prompt Chaining (sequential steps with quality gates), Routing (classify and direct inputs), Parallel Sectioning (independent subtasks in parallel), Parallel Voting (multiple runs for confidence), Orchestrator-Workers (dynamic task decomposition), Evaluator-Optimizer (generate-evaluate loop), Plan-and-Execute (plan then execute with replanning), ReWOO (variable passing between tasks), LLM Compiler (DAG-based parallel execution, 3.6x speedup), and Multi-Agent (collaborative agents with roles).

Orchestrator-Workers vs Plan-and-Execute

Two agent patterns dominate production systems: Plan-and-Execute and Orchestrator-Workers. They solve different problems, and choosing between them correctly is one of the highest-leverage design decisions you will make. Plan-and-Execute is the right pattern when the task structure is predictable. The agent first writes a plan - a fixed sequence of steps - and then executes each step, usually with a separate executor. Because the plan is created up front, the system is easy to inspect, easy to checkpoint, and easy to resume after failures. It works well for structured research: gather sources, extract facts, summarize findings. The weakness is that the plan cannot adapt when the world turns out differently than expected, which forces replanning as a special case. Orchestrator-Workers is the right pattern when the task is dynamic and cannot be divided into a known sequence in advance. A central orchestrator LLM decomposes the task at runtime, dispatches subtasks to worker agents, and integrates their results. Each worker can itself be an agent with its own tools and context. Anthropic recommends this pattern for complex coding tasks, where the shape of the work only becomes clear while the work is being done: a feature might require refactoring, testing, and documentation, and the order depends on what the codebase reveals. The cost of Orchestrator-Workers is control. More agents mean more tokens, more failure modes, more context to manage, and a harder time reproducing results. The orchestrator needs clear instructions for delegation, result validation, and error handling, or the system degrades into many agents confidently doing unrelated work. A practical rule of thumb: if you can write the steps down before the agent starts, use Plan-and-Execute. If the steps only emerge during execution, use Orchestrator-Workers. Many teams build the dynamic version first and pay for complexity they never needed; starting with the simpler pattern and upgrading only when a concrete task demands it produces cheaper, more reliable systems. LoopCraft provides both patterns as visual templates, so you can prototype each against your real task and measure before committing. Available at https://getloopcraft.com.