Setup & Configuration

Create a Pinecone account, install the Python SDK, create your first index, and understand the difference between serverless and pod-based deployments.

Create an Account

1. Sign up at pinecone.io

   Create a free account. The free tier includes 1 serverless index with up to 2GB of storage.

2. Get your API key

   Navigate to "API Keys" in the Pinecone console. Copy your default API key for use in the SDK.

Install the Python SDK

# Install Pinecone client
pip install pinecone

# With gRPC for better performance (optional)
pip install "pinecone[grpc]"

# Set your API key as environment variable
export PINECONE_API_KEY="your-api-key-here"

Initialize the Client

from pinecone import Pinecone, ServerlessSpec
import os

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

# List existing indexes
print(pc.list_indexes())  # []

Creating an Index

An index is where your vectors live. You need to specify the dimensionality (must match your embedding model) and the distance metric:

# Create a serverless index (recommended for most use cases)
pc.create_index(
    name="my-index",
    dimension=1536,        # OpenAI text-embedding-3-small
    metric="cosine",       # cosine, euclidean, or dotproduct
    spec=ServerlessSpec(
        cloud="aws",
        region="us-east-1"
    )
)

# Connect to the index
index = pc.Index("my-index")

# Check index stats
print(index.describe_index_stats())
# {'dimension': 1536, 'total_vector_count': 0, ...}

Serverless vs Pod-Based

Distance Metrics

• cosine: Best for text embeddings (most common). Measures the angle between vectors, ignoring magnitude.
• euclidean: Measures straight-line distance. Use when magnitude matters (e.g., image features).
• dotproduct: Fast computation. Use when vectors are normalized or for recommendation systems.

What's Next?

In the next lesson, we will learn how to upsert vectors into our index, work with metadata, and organize data using namespaces.