langchain-polaris 0.2.0

LangChain integration for The Polaris Report — verified news intelligence tools and retriever

langchain-polaris

LangChain integration for The Polaris Report — verified news intelligence with confidence scores, bias ratings, and source analysis.

Installation

pip install langchain-polaris

Quick Start

Agent with Tools

from langchain_openai import ChatOpenAI
from langchain.agents import AgentExecutor, create_openai_tools_agent
from langchain_core.prompts import ChatPromptTemplate
from langchain_polaris import PolarisSearchTool, PolarisBriefTool, PolarisCompareTool

tools = [
    PolarisSearchTool(api_key="pr_live_xxx"),
    PolarisBriefTool(api_key="pr_live_xxx"),
    PolarisCompareTool(api_key="pr_live_xxx"),
]

prompt = ChatPromptTemplate.from_messages([
    ("system", "You are a news research assistant with access to verified intelligence."),
    ("human", "{input}"),
    ("placeholder", "{agent_scratchpad}"),
])

llm = ChatOpenAI(model="gpt-4o")
agent = create_openai_tools_agent(llm, tools, prompt)
executor = AgentExecutor(agent=agent, tools=tools)

result = executor.invoke({"input": "What's happening with AI regulation?"})
print(result["output"])

Retriever for RAG

from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.runnables import RunnablePassthrough
from langchain_core.output_parsers import StrOutputParser
from langchain_polaris import PolarisRetriever

retriever = PolarisRetriever(
    api_key="pr_live_xxx",
    category="ai_ml",
    min_confidence=0.7,
    limit=5,
)

prompt = ChatPromptTemplate.from_template(
    "Answer based on these verified news briefs:\n\n{context}\n\nQuestion: {question}"
)

chain = (
    {"context": retriever, "question": RunnablePassthrough()}
    | prompt
    | ChatOpenAI(model="gpt-4o")
    | StrOutputParser()
)

print(chain.invoke("What are the latest developments in AI?"))

Tools

Tool	Description
PolarisSearchTool	Search verified news intelligence across 18 verticals. Supports category filtering and speed tiers (depth: fast, standard, deep).
PolarisFeedTool	Get latest verified news briefs, optionally filtered by category or source domain.
PolarisEntityTool	Look up entities (companies, people, technologies) mentioned in verified news coverage.
PolarisBriefTool	Get a specific verified news brief by ID with full analysis, sources, and counter-arguments.
PolarisExtractTool	Extract clean article content from URLs. Returns structured text with metadata.
PolarisVerifyTool	Fact-check a claim against the brief corpus. Returns verdict, confidence, evidence, and nuances. Costs 3 API credits.
PolarisCompareTool	Compare how different news outlets covered the same story. Shows framing, bias, and what each side emphasizes or omits.

Retriever

PolarisRetriever implements LangChain's BaseRetriever interface, returning Document objects with:

• page_content: Headline, summary, and body text
• metadata: brief_id, confidence, bias_score, category, published_at, counter_argument, sources, entities

Retriever Options

Parameter	Type	Default	Description
api_key	str	None	Polaris API key (auto-reads from POLARIS_API_KEY env var or ~/.polaris/credentials if omitted)
category	str	None	Category filter
min_confidence	float	None	Minimum confidence score (0-1)
limit	int	10	Max results to return
include_sources	str	None	Comma-separated source domains to include
exclude_sources	str	None	Comma-separated source domains to exclude

Documentation

Full API docs at thepolarisreport.com/docs