LangChain integration

Add Valyu search and content extraction tools to LangChain and LangGraph agents in Python or TypeScript with langchain-valyu: web, papers, SEC filings and more.

Add Search and Contents to your LangChain or LangGraph application. Your agent controls when to retrieve information and how to use the returned content and source URLs. Choose Python or TypeScript below.

Two tools:

• ValyuSearchTool - search, with full parameter control
• ValyuContentsTool - clean content from specific URLs

Installation

Python

pip install -U langchain-valyu

Set your API key in the environment. The agent examples also want an LLM provider key.

Python

export VALYU_API_KEY="your-valyu-api-key"
export ANTHROPIC_API_KEY="your-anthropic-api-key"
export ANTHROPIC_MODEL="your-supported-model-id"

Search

Python

from langchain_valyu import ValyuSearchTool

tool = ValyuSearchTool()

results = tool._run(
    query="What are agentic search-enhanced large reasoning models?",
    search_type="all",  # "all", "web", or "proprietary"
    max_num_results=5,
    relevance_threshold=0.5,
)

print(results.results)

Content extraction

Clean, structured content from specific URLs, up to 10 per call.

Python

from langchain_valyu import ValyuContentsTool

contents_tool = ValyuContentsTool()

extracted = contents_tool._run(urls=[
    "https://arxiv.org/abs/2301.00001",
    "https://example.com/article",
])

for result in extracted.results:
    print(result["url"], result["title"])
    print(result["content"][:200], "...")

In an agent

The most useful shape: a LangGraph agent that decides for itself when to search.

Python

pip install langchain-anthropic langgraph

Python

import os
from langchain_valyu import ValyuSearchTool
from langchain_anthropic import ChatAnthropic
from langgraph.prebuilt import create_react_agent
from langchain_core.messages import HumanMessage

llm = ChatAnthropic(model=os.environ["ANTHROPIC_MODEL"])
agent = create_react_agent(llm, [ValyuSearchTool()])

question = (
    "What are the key factors driving recent stock market volatility, "
    "and how do macroeconomic indicators influence equity prices across sectors?"
)

for step in agent.stream(
    {"messages": [HumanMessage(content=question)]},
    stream_mode="values",
):
    step["messages"][-1].pretty_print()

Search parameters

The Search tool accepts parameters from the Search API.

Python

results = tool._run(
    query="quantum computing breakthroughs 2024",
    max_num_results=10,          # 1-20; above 20 needs a key permission
    relevance_threshold=0.6,     # 0.0-1.0
    start_date="2024-01-01",     # YYYY-MM-DD
    end_date="2024-12-31",
    included_sources=["valyu/valyu-arxiv", "valyu/valyu-pubmed"],
    response_length="medium",    # "short", "medium", "large", "max", or an int
    country_code="US",
)

Use included_sources or excluded_sources in a request, not both. included_sources takes dataset ids, domains, URL prefixes or presets. ValyuContentsTool takes urls, plus summary, extract_effort and response_length as constructor options.

To steer the agent, bind a system message: use search_type="all" for breadth, "web" for current events only, raise relevance_threshold for precision, cite sources, and use natural-language queries rather than operators. See the Prompting Guide.

Additional resources

LangChain Valyu Tool

Official LangChain integration documentation

API Reference

Complete Valyu API documentation

LangGraph Agents

Build advanced agent workflows

Get API Key

Sign up for $10 free credits ($20 with a company email)