"Text-Embedding-3-Small Complete Guide: OpenAI's Cost-Effective Embedding Model"

from openai import OpenAI

client = OpenAI(
    api_key="your-crazyrouter-key",
    base_url="https://api.crazyrouter.com/v1"
)

# Basic embedding
response = client.embeddings.create(
    model="text-embedding-3-small",
    input="How do I reset my password?"
)

embedding = response.data[0].embedding
print(f"Dimensions: {len(embedding)}")  # 1536
print(f"First 5 values: {embedding[:5]}")

# With dimension reduction
response_small = client.embeddings.create(
    model="text-embedding-3-small",
    input="How do I reset my password?",
    dimensions=512  # Reduce to 512 dimensions
)

embedding_small = response_small.data[0].embedding
print(f"Reduced dimensions: {len(embedding_small)}")  # 512

texts = [
    "How do I reset my password?",
    "I forgot my login credentials",
    "What's the weather like today?",
    "Can I change my email address?",
    "Password recovery not working"
]

response = client.embeddings.create(
    model="text-embedding-3-small",
    input=texts,
    dimensions=512
)

embeddings = [item.embedding for item in response.data]
print(f"Generated {len(embeddings)} embeddings of {len(embeddings[0])} dimensions")

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'your-crazyrouter-key',
  baseURL: 'https://api.crazyrouter.com/v1'
});

async function getEmbedding(text, dimensions = 1536) {
  const response = await client.embeddings.create({
    model: 'text-embedding-3-small',
    input: text,
    dimensions
  });
  return response.data[0].embedding;
}

// Semantic similarity
async function cosineSimilarity(a, b) {
  let dot = 0, normA = 0, normB = 0;
  for (let i = 0; i < a.length; i++) {
    dot += a[i] * b[i];
    normA += a[i] * a[i];
    normB += b[i] * b[i];
  }
  return dot / (Math.sqrt(normA) * Math.sqrt(normB));
}

const emb1 = await getEmbedding('How to reset password');
const emb2 = await getEmbedding('I forgot my login');
const emb3 = await getEmbedding('Best pizza in New York');

console.log('Similar:', await cosineSimilarity(emb1, emb2));  // ~0.85
console.log('Different:', await cosineSimilarity(emb1, emb3)); // ~0.15

curl -X POST https://api.crazyrouter.com/v1/embeddings \
  -H "Authorization: Bearer your-crazyrouter-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": "How do I reset my password?",
    "dimensions": 512
  }'

import numpy as np
from openai import OpenAI

client = OpenAI(
    api_key="your-crazyrouter-key",
    base_url="https://api.crazyrouter.com/v1"
)

DIMENSIONS = 512

# Step 1: Index your documents
documents = [
    "Python is a high-level programming language known for its simplicity.",
    "JavaScript runs in web browsers and powers interactive websites.",
    "Docker containers package applications with their dependencies.",
    "Kubernetes orchestrates container deployments at scale.",
    "PostgreSQL is a powerful open-source relational database.",
    "Redis is an in-memory data store used for caching.",
    "GraphQL is a query language for APIs developed by Facebook.",
    "REST APIs use HTTP methods to perform CRUD operations.",
]

def get_embeddings(texts):
    response = client.embeddings.create(
        model="text-embedding-3-small",
        input=texts,
        dimensions=DIMENSIONS
    )
    return [item.embedding for item in response.data]

# Generate embeddings for all documents
doc_embeddings = np.array(get_embeddings(documents))

# Step 2: Search
def search(query, top_k=3):
    query_embedding = np.array(get_embeddings([query])[0])

    # Cosine similarity
    similarities = np.dot(doc_embeddings, query_embedding) / (
        np.linalg.norm(doc_embeddings, axis=1) * np.linalg.norm(query_embedding)
    )

    top_indices = np.argsort(similarities)[::-1][:top_k]

    results = []
    for idx in top_indices:
        results.append({
            "document": documents[idx],
            "score": float(similarities[idx])
        })
    return results

# Try it
results = search("How do I cache data?")
for r in results:
    print(f"[{r['score']:.3f}] {r['document']}")
# [0.742] Redis is an in-memory data store used for caching.
# [0.385] PostgreSQL is a powerful open-source relational database.
# [0.312] Docker containers package applications with their dependencies.