Search Tool

create_foxnose_tool creates a LangChain tool that wraps FoxNoseRetriever, making FoxNose search available to LLM agents.

Under the hood it uses LangChain's built-in create_retriever_tool, so the resulting tool is fully compatible with any LangChain agent framework.

Quick Start

from foxnose_sdk.flux import FluxClient
from foxnose_sdk.auth import SimpleKeyAuth
from langchain_foxnose import create_foxnose_tool

client = FluxClient(
    base_url="https://<env_key>.fxns.io",
    api_prefix="my_api",
    auth=SimpleKeyAuth("YOUR_PUBLIC_KEY", "YOUR_SECRET_KEY"),
)

tool = create_foxnose_tool(
    client=client,
    folder_path="knowledge-base",
    page_content_field="body",
)

Using with an Agent

from langchain_openai import ChatOpenAI
from langgraph.prebuilt import create_react_agent

llm = ChatOpenAI(model="gpt-4o")
agent = create_react_agent(llm, tools=[tool])

result = agent.invoke({"messages": [{"role": "user", "content": "How do I reset my password?"}]})

Wrapping an Existing Retriever

If you already have a configured FoxNoseRetriever, pass it directly:

from langchain_foxnose import FoxNoseRetriever, create_foxnose_tool

retriever = FoxNoseRetriever(
    client=client,
    folder_path="knowledge-base",
    page_content_field="body",
    search_mode="hybrid",
    top_k=5,
)

tool = create_foxnose_tool(
    retriever=retriever,
    name="kb_search",
    description="Search the internal knowledge base for answers.",
)

Parameters

Parameter	Default	Description
client	None	Sync FluxClient (used to build a retriever if none provided)
async_client	None	Async AsyncFluxClient
retriever	None	Existing FoxNoseRetriever to wrap
folder_path	—	Folder path (required when building a new retriever)
page_content_field	None	Data field for page_content
name	"foxnose_search"	Tool name exposed to the agent
description	"Search the FoxNose…"	Tool description exposed to the agent
document_separator	"\n\n"	Separator between documents in the response
response_format	"content"	"content" for plain text, "content_and_artifact" for (str, list[Document])
**retriever_kwargs	—	Extra args forwarded to FoxNoseRetriever

Response Formats

• "content" (default) — The tool returns a plain string with document contents joined by document_separator. Best for simple agents.
• "content_and_artifact" — The tool returns a tuple (str, list[Document]). Use this when your agent needs access to raw Document objects and their metadata.

API Reference

langchain_foxnose.tools.create_foxnose_tool(*, client=None, async_client=None, retriever=None, folder_path=None, page_content_field=None, name='foxnose_search', description='Search the FoxNose knowledge base. Use this tool to find relevant information for answering questions.', document_separator='\n\n', response_format='content', **retriever_kwargs)

Create a LangChain tool for FoxNose search.

Either pass an existing retriever or provide client + config to create one internally.

Example

.. code-block:: python

from foxnose_sdk.flux import FluxClient
from foxnose_sdk.auth import SimpleKeyAuth
from langchain_foxnose import create_foxnose_tool

client = FluxClient(
    base_url="https://<env_key>.fxns.io",
    api_prefix="my_api",
    auth=SimpleKeyAuth("pk", "sk"),
)
tool = create_foxnose_tool(
    client=client,
    folder_path="knowledge-base",
    page_content_field="body",
)

Parameters:

Name	Type	Description	Default
client	Any | None	Synchronous :class:~foxnose_sdk.flux.FluxClient instance.	None
async_client	Any | None	Asynchronous :class:~foxnose_sdk.flux.AsyncFluxClient instance.	None
retriever	FoxNoseRetriever | None	An existing :class:FoxNoseRetriever to wrap. If provided, client, async_client, folder_path, page_content_field, and retriever_kwargs are ignored.	None
folder_path	str | None	Folder path in FoxNose (required when building a new retriever).	None
page_content_field	str | None	Single data field for page_content (required when building a new retriever, unless another content strategy is provided via retriever_kwargs).	None
name	str	Tool name exposed to the LLM agent.	'foxnose_search'
description	str	Tool description exposed to the LLM agent.	'Search the FoxNose knowledge base. Use this tool to find relevant information for answering questions.'
document_separator	str	Separator between documents in the response string.	'\n\n'
response_format	Literal['content', 'content_and_artifact']	"content" returns a plain string; "content_and_artifact" returns (str, list[Document]).	'content'
**retriever_kwargs	Any	Extra keyword arguments forwarded to :class:FoxNoseRetriever when creating a new instance.	{}

Returns:

Name	Type	Description
A	BaseTool	class:~langchain_core.tools.BaseTool that performs FoxNose search.

Raises:

Type	Description
ValueError	If neither retriever nor client/async_client is provided.