How to beginner · 3 min read

How to create a Pinecone index

Q: How to create a Pinecone index

Use the Pinecone SDK v3+ to create an index by initializing the client with your API key, then calling pc.Index.create() with your index name, dimension, and metric. This sets up a vector index ready for storing and querying embeddings.

Quick answer

Use the Pinecone SDK v3+ to create an index by initializing the client with your API key, then calling pc.Index.create() with your index name, dimension, and metric. This sets up a vector index ready for storing and querying embeddings.

PREREQUISITES

Python 3.8+
Pinecone API key
pip install pinecone-client>=3.0

Setup

Install the Pinecone Python SDK and set your API key as an environment variable before creating an index.

bash

pip install pinecone-client>=3.0

Step by step

This example shows how to create a Pinecone index named example-index with 1536 dimensions and cosine similarity metric.

python

import os
from pinecone import Pinecone

# Initialize Pinecone client
pc = Pinecone(api_key=os.environ["PINECONE_API_KEY"])

# Define index parameters
index_name = "example-index"
dimension = 1536
metric = "cosine"

# Create the index
pc.Index.create(name=index_name, dimension=dimension, metric=metric)

print(f"Index '{index_name}' created successfully.")

output

Index 'example-index' created successfully.

Common variations

You can create indexes with different metrics like euclidean or dotproduct, and adjust the dimension to match your embedding size. The Pinecone SDK also supports async usage with asyncio if needed.

python

import asyncio
from pinecone import Pinecone

async def create_index_async():
    pc = Pinecone(api_key=os.environ["PINECONE_API_KEY"])
    await pc.Index.create(name="async-index", dimension=512, metric="euclidean")
    print("Async index created.")

asyncio.run(create_index_async())

output

Async index created.

Troubleshooting

If you see IndexAlreadyExistsException, the index name is taken; choose a unique name.
Ensure your PINECONE_API_KEY environment variable is set correctly.
Check your network connection if the client cannot reach Pinecone servers.

✅

Key Takeaways

Initialize the Pinecone client with your API key from environment variables.
Use pc.Index.create() with name, dimension, and metric to create an index.
Match the dimension to your embedding vector size for proper indexing.
Choose the similarity metric based on your use case: cosine, euclidean, or dotproduct.
Handle common errors like duplicate index names and API key misconfiguration.

Verified 2026-04

Verify ↗