Overview

Use case: General retrieval, code search, multilingual, etc.
Features needed: input_type, dimensions
Cost: varies significantly by provider
Performance: latency and throughput requirements
Context length: maximum input tokens

Catsu supports 11 major embedding providers with a unified, consistent API.

Supported Providers

Provider	Models	Key Features	Environment Variable
OpenAI	3	Industry standard, Matryoshka	`OPENAI_API_KEY`
Voyage AI	11	Domain-specific, multimodal, quantization	`VOYAGE_API_KEY`
Cohere	5	Multilingual, truncation	`COHERE_API_KEY`
Gemini	1	Long context, Matryoshka	`GEMINI_API_KEY`
Jina AI	6	Multimodal, code-specific, long context	`JINA_API_KEY`
Mistral AI	2	Code-optimized, quantization	`MISTRAL_API_KEY`
Nomic	2	Long text handling, Matryoshka	`NOMIC_API_KEY`
Cloudflare	7	Edge inference, BGE models	`CLOUDFLARE_API_TOKEN`
DeepInfra	16	Open-source models, Qwen3	`DEEPINFRA_API_KEY`
Mixedbread	4	Multilingual, quantization	`MIXEDBREAD_API_KEY`
Together AI	7	Open-source, long context	`TOGETHER_API_KEY`

For detailed model information including pricing and benchmarks, visit the Models Catalog.

Quick Examples

Using Different Providers

from catsu import Client

client = Client()

# OpenAI
openai_response = client.embed("openai:text-embedding-3-small", ["Text"])

# Voyage AI
voyage_response = client.embed("voyageai:voyage-3", ["Text"])

# Cohere
cohere_response = client.embed("cohere:embed-v4.0", ["Text"])

Provider-Specific Features

# OpenAI with custom dimensions
response = client.embed(
    "openai:text-embedding-3-small",
    ["Text"],
    dimensions=256
)

# Voyage AI with input_type
response = client.embed(
    "voyageai:voyage-3",
    ["Search query"],
    input_type="query"
)

# Gemini with long context (2048 tokens)
response = client.embed(
    "gemini:gemini-embedding-001",
    ["Very long document..."]
)