How to beginner · 3 min read

How to get started with Weaviate

Q: How to get started with Weaviate

To get started with Weaviate, install the weaviate-client Python package and connect to a running Weaviate instance. Use the client to create schemas, add vectorized data, and perform semantic searches with simple Python API calls.

Quick answer

To get started with Weaviate, install the weaviate-client Python package and connect to a running Weaviate instance. Use the client to create schemas, add vectorized data, and perform semantic searches with simple Python API calls.

PREREQUISITES

Python 3.8+
Running Weaviate instance (local Docker or cloud)
pip install weaviate-client

Setup

Install the official weaviate-client Python package and ensure you have access to a running Weaviate server, either locally via Docker or a cloud instance.

bash

pip install weaviate-client

Step by step

This example connects to a local Weaviate instance, creates a schema, adds an object with a vector, and queries it.

python

import weaviate

# Connect to local Weaviate instance
client = weaviate.Client("http://localhost:8080")

# Define a simple schema
schema = {
    "classes": [
        {
            "class": "Article",
            "description": "A news article",
            "properties": [
                {
                    "name": "title",
                    "dataType": ["text"]
                },
                {
                    "name": "content",
                    "dataType": ["text"]
                }
            ]
        }
    ]
}

# Create schema
client.schema.delete_all()  # Clear existing schema
client.schema.create(schema)

# Add an object with vector
obj = {
    "title": "Weaviate introduction",
    "content": "Weaviate is an open-source vector search engine."
}

# Example vector (usually from an embedding model)
vector = [0.1, 0.2, 0.3, 0.4]

client.data_object.create(obj, "Article", vector=vector)

# Query by vector (search for similar vectors)
near_vector = {"vector": [0.1, 0.2, 0.3, 0.4], "certainty": 0.7}
result = client.query.get("Article", ["title", "content"]).with_near_vector(near_vector).do()

print(result)

output

{'data': {'Get': {'Article': [{'title': 'Weaviate introduction', 'content': 'Weaviate is an open-source vector search engine.'}]}}}

Common variations

You can connect to cloud-hosted Weaviate instances by changing the URL and adding authentication headers. Async usage is not supported in the official client. Use different embedding models to generate vectors before adding data.

python

import weaviate

# Connect to cloud Weaviate with API key
client = weaviate.Client(
    url="https://your-cloud-instance.weaviate.network",
    auth_client_secret=weaviate.AuthApiKey(api_key=os.environ["OPENAI_API_KEY"])
)

# Use client as usual for schema and data operations

Troubleshooting

If you get connection errors, verify your Weaviate server is running and accessible at the specified URL.
Schema creation errors often mean the schema already exists; use client.schema.delete_all() to reset.
Ensure your vectors match the dimensionality expected by your Weaviate instance.

✅

Key Takeaways

Install weaviate-client and connect to a running Weaviate server to start.
Define schemas and add vectorized data objects via the Python client API.
Use with_near_vector queries to perform semantic vector searches.
Cloud instances require API key authentication with AuthApiKey.
Clear schema conflicts with client.schema.delete_all() before creating new schemas.

Verified 2026-04

Verify ↗