How to Intermediate · 4 min read

How to build Perplexity-like search

Q: How to build Perplexity-like search

Build a Perplexity-like search by combining OpenAI embeddings to vectorize documents, a vector database like FAISS or Chroma for similarity search, and OpenAI chat completions to generate natural language answers from retrieved context. This retrieval-augmented generation (RAG) approach enables precise, conversational search results.

Quick answer

Build a Perplexity-like search by combining OpenAI embeddings to vectorize documents, a vector database like FAISS or Chroma for similarity search, and OpenAI chat completions to generate natural language answers from retrieved context. This retrieval-augmented generation (RAG) approach enables precise, conversational search results.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0 langchain-openai langchain-community faiss-cpu

Setup

Install required Python packages and set your OPENAI_API_KEY environment variable.

Use pip install openai langchain-openai langchain-community faiss-cpu
Export your API key: export OPENAI_API_KEY='your_api_key' (Linux/macOS) or set in your environment on Windows.

bash

pip install openai langchain-openai langchain-community faiss-cpu

output

Collecting openai
Collecting langchain-openai
Collecting langchain-community
Collecting faiss-cpu
Successfully installed openai langchain-openai langchain-community faiss-cpu

Step by step

This example loads documents, creates embeddings with OpenAIEmbeddings, indexes them in FAISS, performs similarity search, and uses OpenAI chat completions to answer queries based on retrieved context.

python

import os
from openai import OpenAI
from langchain_openai import OpenAIEmbeddings
from langchain_community.vectorstores import FAISS
from langchain_core.prompts import ChatPromptTemplate

# Initialize OpenAI client
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Sample documents to index
documents = [
    "Python is a popular programming language.",
    "OpenAI provides powerful AI models.",
    "FAISS is a library for efficient similarity search.",
    "Perplexity AI combines search with chat completions."
]

# Create embeddings for documents
embeddings = OpenAIEmbeddings(client=client, model="text-embedding-3-small")

# Build vector store index
vector_store = FAISS.from_texts(documents, embeddings)

# User query
query = "What is FAISS used for?"

# Embed query and retrieve top 2 similar docs
docs = vector_store.similarity_search(query, k=2)

# Prepare context from retrieved docs
context = "\n".join([doc.page_content for doc in docs])

# Define prompt template with context
prompt_template = ChatPromptTemplate.from_template(
    """
You are a helpful assistant. Use the following context to answer the question.
Context:\n{context}
Question: {question}
Answer:
"""
)

# Format prompt
messages = [
    {"role": "user", "content": prompt_template.format(context=context, question=query)}
]

# Generate answer using chat completion
response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=messages
)

print("Answer:", response.choices[0].message.content)

output

Answer: FAISS is a library for efficient similarity search, commonly used to quickly find similar vectors in large datasets.

Common variations

You can enhance your Perplexity-like search by:

Using Chroma or other vector stores instead of FAISS.
Switching to a larger model like gpt-4o for more detailed answers.
Implementing async calls with asyncio and await for better performance.
Adding streaming responses for real-time answer generation.

python

import asyncio
from openai import OpenAI

async def async_search():
    client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
    # Async chat completion example
    response = await client.chat.completions.acreate(
        model="gpt-4o",
        messages=[{"role": "user", "content": "Explain vector search."}]
    )
    print("Async answer:", response.choices[0].message.content)

asyncio.run(async_search())

output

Async answer: Vector search is a technique to find similar items by comparing their vector embeddings, enabling fast and accurate retrieval in large datasets.

Troubleshooting

If you get InvalidRequestError, verify your OPENAI_API_KEY is set correctly.
If similarity search returns no results, check that embeddings are generated and indexed properly.
For slow responses, consider using smaller models or caching embeddings.
Ensure your documents are clean and relevant for better retrieval quality.

✅

Key Takeaways

Use OpenAI embeddings with a vector store like FAISS to enable semantic search.
Combine retrieved documents as context for chat completions to generate natural answers.
Leverage LangChain's OpenAIEmbeddings and vectorstore integrations for rapid development.

Verified 2026-04 · gpt-4o-mini, text-embedding-3-small

Verify ↗