LangChain

pip install -U langchain langchain-nimble langchain-openai

export NIMBLE_API_KEY="your-api-key"

from langchain.agents import create_agent
from langchain_nimble import NimbleToolkit
from langchain_openai import ChatOpenAI

# Initialize toolkit — search and extract are enabled by default
toolkit = NimbleToolkit(
    include_crawl=True,
    include_map=True,
    include_agent=True,
)
tools = toolkit.get_tools()

# Create agent with all Nimble tools
agent = create_agent(
    model=ChatOpenAI(model="gpt-5"),
    tools=tools,
    system_prompt=(
        "You are a helpful research assistant with access to "
        "real-time web information. You can search the web, "
        "extract content from URLs, crawl entire sites, "
        "map site structures, and use specialized data agents."
    )
)

# Use the agent
response = agent.invoke({
    "messages": [(
        "user",
        "What are the latest developments in AI agents? "
        "Summarize key findings."
    )]
})

print(response["messages"][-1].content)

from langchain_nimble import NimbleCrawlTool

crawl_tool = NimbleCrawlTool()

result = crawl_tool.invoke({
    "url": "https://docs.example.com",
    "max_pages": 50
})

from langchain_nimble import NimbleMapTool

map_tool = NimbleMapTool()

result = map_tool.invoke({
    "url": "https://www.example.com"
})

from langchain_nimble import NimbleToolkit

toolkit = NimbleToolkit(include_agent=True)
tools = toolkit.get_tools()
# Includes: nimble_agent_list, nimble_agent_get, nimble_agent_run

from langchain_nimble import NimbleExtractRetriever
from langchain_openai import ChatOpenAI
from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.runnables import RunnablePassthrough

# Extract your knowledge base from specific URLs
retriever = NimbleExtractRetriever(
    urls=[
        "https://python.langchain.com/docs/concepts/retrievers/",
        "https://python.langchain.com/docs/concepts/tools/",
        "https://python.langchain.com/docs/tutorials/agents/"
    ],
    output_format="markdown"
)

# Build RAG chain
llm = ChatOpenAI(model="gpt-5", temperature=0)
prompt = ChatPromptTemplate.from_template(
    """Answer based only on the provided documentation.

Documentation: {context}

Question: {question}"""
)

chain = (
    {"context": retriever | (lambda docs: "\n\n".join(doc.page_content for doc in docs)),
     "question": RunnablePassthrough()}
    | prompt
    | llm
    | StrOutputParser()
)

# Ask questions about your knowledge base
answer = chain.invoke("What are the key differences between retrievers and tools in LangChain?")
print(answer)