Additional Frameworks & Quick Reference

LangChain

Best for: All-in-one orchestration, chains and RAG

from langchain.agents import AgentExecutor, create_tool_using_agent

agent = create_tool_using_agent(
 llm,
 tools,
 agent="zero-shot-react-description"
)

executor = AgentExecutor.from_agent_and_tools(
 agent=agent,
 tools=tools,
 verbose=True
)

result = executor.run("Find information about X")

-

Semantic Kernel (Microsoft)

Best for: Enterprise integration, C#/.NET

var builder = new KernelBuilder();
var kernel = builder
.WithAzureOpenAIChatCompletion(...)
.Build();

var response = await kernel.RunAsync(
 "Find research about topic",
 new AIRequestSettings()
);

---

Haystack

Best for: Search and retrieval pipelines

from haystack import Pipeline, Document
from haystack.components.retrievers import InMemoryBM25Retriever

pipeline = Pipeline()
pipeline.add_component("retriever", InMemoryBM25Retriever())
pipeline.connect("prompt", "retriever")

result = pipeline.run({"retriever": {"query": "What is AI?"}})

---

Framework Comparison Matrix

Framework	Ease	Power	Production	Community
Claude API				Growing
LangGraph				Very Active
CrewAI				Active
AutoGen				Very Active
LangChain				Very Active

-

Last Updated: August 9, 2026