LangChain

Retrieval chains stuff whole documents into the prompt. That is slow and costly, and a chunk cut from its page loses who and when it was about. Flux gives your agent live search and trims what it retrieves to the passages that matter, with the names and dates written back in.

Build it

1. 1

   Install

   Shell

   pip install langchain langchain-openai requestsexport FLUX_API_KEY="<your key>"

2. 2

   Give your agent web search

   Your model plans and answers. Flux is the tool it calls for anything current.

   Python

   import osimport requestsFLUX = "https://fluxsearch.io/api/v1"HEADERS = {"Authorization": f"Bearer {os.environ['FLUX_API_KEY']}"}from langchain.agents import create_agentfrom langchain_core.tools import tool@tooldef web_search(query: str) -> str:    """Search the live web. Returns passages with provenance and their source URLs."""    response = requests.post(        f"{FLUX}/search",        headers=HEADERS,        json={"query": query, "max_results": 5, "enrich": 3},        timeout=120,    )    response.raise_for_status()    passages = [        f"{passage['enriched_text']} ({result['url']})"        for result in response.json()["results"]        for passage in (result["lineage"] or {}).get("results", [])    ]    return "\n\n".join(passages) or "No passages found."agent = create_agent("openai:gpt-5-mini", tools=[web_search])result = agent.invoke({"messages": [{"role": "user", "content": "Who runs Barclays?"}]})print(result["messages"][-1].content)

   Show all 31 lines

3. 3

   Trim retrieved documents

   Wrap Lineage as a BaseDocumentCompressor. Each document comes back as the passages that answer the query.

   Python

   import osimport requestsFLUX = "https://fluxsearch.io/api/v1"HEADERS = {"Authorization": f"Bearer {os.environ['FLUX_API_KEY']}"}from langchain_core.documents import Documentfrom langchain_core.documents.compressor import BaseDocumentCompressorclass FluxLineage(BaseDocumentCompressor):    def compress_documents(self, documents, query, callbacks=None):        kept = []        for doc in documents:            response = requests.post(                f"{FLUX}/provenance",                headers=HEADERS,                json={"text": doc.page_content, "query": query},                timeout=120,            )            response.raise_for_status()            for passage in response.json()["results"]:                metadata = {**doc.metadata, "score": passage["score"]}                kept.append(Document(page_content=passage["enriched_text"], metadata=metadata))        return keptdocs = FluxLineage().compress_documents(retrieved_docs, "who is the chief executive")

   Show all 28 lines

Call Flux as a chat model

For a direct answer with no agent, ChatOpenAI works against Flux with a new base URL.

import osfrom langchain_openai import ChatOpenAIflux = ChatOpenAI(    model="flux-search-1",    base_url="https://fluxsearch.io/api/v1",    api_key=os.environ["FLUX_API_KEY"],)print(flux.invoke("Who is the chief executive of Barclays?").content)

Next steps