Gemini Embeddings API - Generate text-embedding vectors with Google's embedding models. Build semantic search, RAG, and similarity matching applications.

Embed Content

Generate vector embeddings for text content. Embeddings are numerical representations of text that capture semantic meaning, useful for search, clustering, and similarity comparisons.

Endpoint

POST https://api.uniterouter.ai/v1beta/models/{model}:embedContent

Authentication

x-goog-api-key: your_uniterouter_api_key

Path Parameters

Parameter	Type	Required	Description
`model`	string	Yes	Embedding model ID (e.g., `text-embedding-004`)

Request Body

Parameter	Type	Required	Description
`content`	object	Yes	Content to embed
`taskType`	string	No	Type of task for the embedding
`title`	string	No	Optional title for the content
`outputDimensionality`	integer	No	Desired embedding dimensions

Task Types

Task Type	Description
`RETRIEVAL_QUERY`	For search queries
`RETRIEVAL_DOCUMENT`	For documents to be searched
`SEMANTIC_SIMILARITY`	For comparing text similarity
`CLASSIFICATION`	For text classification
`CLUSTERING`	For clustering text

Example Request

curl -X POST "https://api.uniterouter.ai/v1beta/models/text-embedding-004:embedContent" \
  -H "Content-Type: application/json" \
  -H "x-goog-api-key: $UNITEROUTER_API_KEY" \
  -d '{
    "content": {
      "parts": [{"text": "What is machine learning?"}]
    },
    "taskType": "RETRIEVAL_QUERY"
  }'

import google.generativeai as genai

genai.configure(
    api_key="your_uniterouter_api_key",
    transport="rest",
    client_options={"api_endpoint": "api.uniterouter.ai"}
)

result = genai.embed_content(
    model="models/text-embedding-004",
    content="What is machine learning?",
    task_type="retrieval_query"
)

print(f"Embedding dimension: {len(result['embedding'])}")
print(f"First 5 values: {result['embedding'][:5]}")

import { GoogleGenerativeAI } from "@google/generative-ai";

const genAI = new GoogleGenerativeAI("your_uniterouter_api_key", {
  baseUrl: "https://api.uniterouter.ai"
});

const model = genAI.getGenerativeModel({ model: "text-embedding-004" });

const result = await model.embedContent({
  content: { parts: [{ text: "What is machine learning?" }] },
  taskType: "RETRIEVAL_QUERY"
});

console.log(`Embedding dimension: ${result.embedding.values.length}`);

Response

{
  "embedding": {
    "values": [0.0123, -0.0456, 0.0789, ...]
  }
}

The values array contains the embedding vector (typically 768 dimensions for text-embedding-004).

Batch Embeddings

For multiple texts, use the batch endpoint:

POST /v1beta/models/{model}:batchEmbedContents

texts = [
    "What is machine learning?",
    "How does deep learning work?",
    "Explain neural networks"
]

results = genai.embed_content(
    model="models/text-embedding-004",
    content=texts,
    task_type="retrieval_document"
)

for i, embedding in enumerate(results['embedding']):
    print(f"Text {i}: {len(embedding)} dimensions")

Embedding Models

Model	Dimensions	Description
`text-embedding-004`	768	Latest embedding model
`embedding-001`	768	Previous generation

Use Cases

Semantic Search

import numpy as np

def cosine_similarity(a, b):
    return np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b))

# Embed query
query_embedding = genai.embed_content(
    model="models/text-embedding-004",
    content="How to train a model?",
    task_type="retrieval_query"
)['embedding']

# Embed documents
documents = ["Training ML models...", "Data preprocessing...", "Model evaluation..."]
doc_embeddings = genai.embed_content(
    model="models/text-embedding-004",
    content=documents,
    task_type="retrieval_document"
)['embedding']

# Find most similar
similarities = [cosine_similarity(query_embedding, doc) for doc in doc_embeddings]
best_match = documents[np.argmax(similarities)]

RAG (Retrieval-Augmented Generation)

# 1. Embed and store documents
# 2. Embed user query
# 3. Find relevant documents
# 4. Generate response with context

query = "What are transformers?"
query_embed = genai.embed_content(
    model="models/text-embedding-004",
    content=query,
    task_type="retrieval_query"
)['embedding']

# Retrieve relevant docs from vector database
relevant_docs = vector_db.search(query_embed, top_k=3)

# Generate with context
model = genai.GenerativeModel("gemini-2.0-flash")
response = model.generate_content(f"""
Context: {relevant_docs}

Question: {query}

Answer based on the context:
""")

Generate Content - Text generation
Embeddings (OpenAI) - OpenAI-compatible embeddings

Embed Content

On this page