LangChain RAG Search Grounding Workflow

from langchain.schema.runnable import RunnableLambda, RunnablePassthrough
from langchain_core.prompts import ChatPromptTemplate
from langchain_openai import ChatOpenAI
from langchain_core.output_parsers import StrOutputParser
import requests

SCRAVIO_KEY = "YOUR_API_KEY"

def scavio_retriever(query: str) -> str:
    """Fetch search results and format as context string."""
    resp = requests.post(
        "https://api.scavio.dev/api/v1/search",
        headers={"x-api-key": SCRAVIO_KEY},
        json={"query": query, "platform": "google", "num": 5}
    )
    resp.raise_for_status()
    results = resp.json().get("results", [])
    context_parts = []
    for i, r in enumerate(results[:5], start=1):
        title = r.get("title", "")
        snippet = r.get("snippet", "")
        url = r.get("url", "")
        context_parts.append(f"[Source {i}: {url}]\n{title}\n{snippet}")
    return "\n\n".join(context_parts)[:3000]

SYSTEM_PROMPT = """You are a helpful assistant that answers questions using only the provided search results.
Always cite your sources using [Source N: URL] format.
If the search results don't contain enough information, say so — do not fabricate facts."""

prompt = ChatPromptTemplate.from_messages([
    ("system", SYSTEM_PROMPT),
    ("human", "Search results:\n{context}\n\nQuestion: {question}")
])

llm = ChatOpenAI(model="gpt-4o-mini", temperature=0.1)

# Build the chain
chain = (
    RunnablePassthrough.assign(
        context=RunnableLambda(lambda x: scavio_retriever(x["question"]))
    )
    | prompt
    | llm
    | StrOutputParser()
)

def grounded_answer(question: str) -> dict:
    answer = chain.invoke({"question": question})
    return {"question": question, "answer": answer}

if __name__ == "__main__":
    result = grounded_answer("What does Scavio cost per month?")
    print(result["answer"])

const { ChatOpenAI } = require('@langchain/openai');
const { ChatPromptTemplate } = require('@langchain/core/prompts');
const { StringOutputParser } = require('@langchain/core/output_parsers');
const { RunnablePassthrough, RunnableLambda } = require('@langchain/core/runnables');
const fetch = require('node-fetch');

const SCRAVIO_KEY = 'YOUR_API_KEY';

async function scavioRetriever(query) {
  const res = await fetch('https://api.scavio.dev/api/v1/search', {
    method: 'POST',
    headers: { 'x-api-key': SCRAVIO_KEY, 'Content-Type': 'application/json' },
    body: JSON.stringify({ query, platform: 'google', num: 5 })
  });
  const results = (await res.json()).results || [];
  return results.slice(0, 5).map((r, i) =>
    `[Source ${i+1}: ${r.url}]\n${r.title}\n${r.snippet}`
  ).join('\n\n').slice(0, 3000);
}

const SYSTEM = 'Answer using only the provided search results. Cite sources as [Source N: URL]. Do not fabricate facts.';

const prompt = ChatPromptTemplate.fromMessages([
  ['system', SYSTEM],
  ['human', 'Search results:\n{context}\n\nQuestion: {question}']
]);

const llm = new ChatOpenAI({ model: 'gpt-4o-mini', temperature: 0.1 });

const chain = RunnablePassthrough.assign({
  context: new RunnableLambda({ func: async (x) => scavioRetriever(x.question) })
}).pipe(prompt).pipe(llm).pipe(new StringOutputParser());

async function groundedAnswer(question) {
  const answer = await chain.invoke({ question });
  return { question, answer };
}

groundedAnswer('What does Scavio cost per month?').then(r => console.log(r.answer));