Embedding Providers

Generate text embeddings using 35+ providers through PraisonAI

Overview

PraisonAI supports 35+ embedding providers through LiteLLM integration, giving you access to hundreds of embedding models with a unified API.

Quick Start

from praisonaiagents import embedding

# OpenAI (default)
result = embedding("Hello world", model="text-embedding-3-small")

# Cohere
result = embedding("Hello world", model="cohere/embed-english-v3.0")

# Voyage AI
result = embedding("Hello world", model="voyage/voyage-3")

# Azure OpenAI
result = embedding("Hello world", model="azure/text-embedding-ada-002")

print(f"Dimensions: {len(result.embeddings[0])}")

CLI Usage

# OpenAI
praisonai embed "Hello world" --model text-embedding-3-small

# Cohere
praisonai embed "Hello world" --model cohere/embed-english-v3.0

# Any provider
praisonai embedding "Hello world" --model voyage/voyage-3

Supported Providers

Tier 1: Major Cloud Providers

Provider	Prefix	Example Model	Docs
OpenAI	openai/ or none	text-embedding-3-small	→
Azure OpenAI	azure/	azure/text-embedding-ada-002	→
Google Vertex AI	vertex_ai/	vertex_ai/textembedding-gecko	→
Google Gemini	gemini/	gemini/text-embedding-004	→
AWS Bedrock	bedrock/	bedrock/amazon.titan-embed-text-v1	→
Azure AI	azure_ai/	azure_ai/Cohere-embed-v3-english	→

Tier 2: Specialized Embedding Providers

Provider	Prefix	Example Model	Docs
Cohere	cohere/	cohere/embed-english-v3.0	→
Voyage AI	voyage/	voyage/voyage-3	→
Jina AI	jina_ai/	jina_ai/jina-embeddings-v3	→
Mistral	mistral/	mistral/mistral-embed	→

Tier 3: Open Source & Self-Hosted

Provider	Prefix	Example Model	Docs
HuggingFace	huggingface/	huggingface/BAAI/bge-large-en-v1.5	→
Ollama	ollama/	ollama/nomic-embed-text	→
Infinity	infinity/	infinity/BAAI/bge-small-en-v1.5	→
vLLM	hosted_vllm/	hosted_vllm/intfloat/e5-mistral-7b-instruct	→
LM Studio	lm_studio/	lm_studio/nomic-embed-text	→

Tier 4: Additional Providers

Provider	Prefix	Example Model	Docs
Together AI	together_ai/	together_ai/togethercomputer/m2-bert-80M-8k-retrieval	→
Fireworks AI	fireworks_ai/	fireworks_ai/nomic-ai/nomic-embed-text-v1.5	→
NVIDIA NIM	nvidia_nim/	nvidia_nim/NV-Embed-QA	→
Databricks	databricks/	databricks/databricks-bge-large-en	→
Snowflake	snowflake/	snowflake/snowflake-arctic-embed-m	→
Watsonx	watsonx/	watsonx/ibm/slate-125m-english-rtrvr	→
SambaNova	sambanova/	sambanova/E5-mistral-7b-instruct	→
Nebius	nebius/	nebius/BAAI/bge-en-icl	→
OVHcloud	ovhcloud/	ovhcloud/multilingual-e5-base	→
Volcengine	volcengine/	volcengine/doubao-embedding	→

Model Selection Guide

By Use Case

Use Case	Recommended Model	Dimensions
General Purpose	text-embedding-3-small	1536
High Quality	text-embedding-3-large	3072
Multilingual	cohere/embed-multilingual-v3.0	1024
Code Search	voyage/voyage-code-3	1024
Cost Effective	cohere/embed-english-light-v3.0	384
Self-Hosted	ollama/nomic-embed-text	768

By Dimension Size

Dimensions	Models
256-512	cohere/embed-english-light-v3.0, jina_ai/jina-embeddings-v2-small-en
768-1024	cohere/embed-english-v3.0, voyage/voyage-3, mistral/mistral-embed
1536	text-embedding-3-small, azure/text-embedding-ada-002
3072	text-embedding-3-large

Environment Variables

Each provider requires specific environment variables:

# OpenAI
export OPENAI_API_KEY="sk-..."

# Azure OpenAI
export AZURE_API_KEY="..."
export AZURE_API_BASE="https://your-resource.openai.azure.com"

# Cohere
export COHERE_API_KEY="..."

# Voyage AI
export VOYAGE_API_KEY="..."

# Google
export GOOGLE_API_KEY="..."  # or GEMINI_API_KEY

# AWS Bedrock
export AWS_ACCESS_KEY_ID="..."
export AWS_SECRET_ACCESS_KEY="..."
export AWS_REGION_NAME="us-east-1"

API Reference

Parameters

Parameter	Type	Default	Description
input	str or List[str]	Required	Text(s) to embed
model	str	text-embedding-3-small	Model identifier with optional provider prefix
dimensions	int	None	Output dimensions (if supported)
encoding_format	str	float	float or base64
timeout	float	600.0	Request timeout in seconds
api_key	str	None	API key override
api_base	str	None	API base URL override

Response

EmbeddingResult(
    embeddings=[[0.1, 0.2, ...]],  # List of embedding vectors
    model="text-embedding-3-small",
    usage={"prompt_tokens": 4, "total_tokens": 4}
)

Related

• Embeddings API - Core embeddings documentation
• Embeddings CLI - CLI commands
• Vector Stores - Store and query embeddings