Supported Providers

Proxle supports 5 major LLM providers. Each provider can be used through the SDK wrappers or by calling the proxy directly.

OpenAI

SDK wrapper: OpenAI / AsyncOpenAI

Supported endpoints:

• /chat/completions - Chat completions (GPT-4o, GPT-4, GPT-3.5, O1, O3)
• /embeddings - Text embeddings

Setup:

from proxle import OpenAI

client = OpenAI(
    api_key="sk-...",
    proxy_key="pk_live_..."
)

Proxy URL: POST /v1/proxy/openai/chat/completions

Anthropic

SDK wrapper: Anthropic / AsyncAnthropic

Supported endpoints:

• /messages - Messages API (Claude Opus, Sonnet, Haiku)

Setup:

from proxle import Anthropic

client = Anthropic(
    api_key="sk-ant-...",
    proxy_key="pk_live_..."
)

Proxy URL: POST /v1/proxy/anthropic/messages

Cohere

SDK wrapper: Cohere (lightweight, no official SDK required)

Supported endpoints:

• /chat - Chat completions (Command R+, Command R)
• /generate - Text generation
• /embed - Text embeddings

Setup:

from proxle import Cohere

client = Cohere(
    api_key="...",
    proxy_key="pk_live_..."
)

Proxy URL: POST /v1/proxy/cohere/chat

Google Gemini

SDK wrapper: Gemini (lightweight, no official SDK required)

Supported endpoints:

• /generateContent - Content generation (Gemini 2.0 Flash, 1.5 Pro, 1.5 Flash)
• /embedContent - Text embeddings
• /countTokens - Token counting

Setup:

from proxle import Gemini

client = Gemini(
    api_key="...",
    proxy_key="pk_live_..."
)

Proxy URL: POST /v1/proxy/gemini/models/{model}:generateContent

Azure OpenAI

SDK wrapper: AzureOpenAI / AsyncAzureOpenAI

Supported endpoints:

• /chat/completions - Chat completions
• /embeddings - Text embeddings

Setup:

from proxle import AzureOpenAI

client = AzureOpenAI(
    api_key="...",
    azure_endpoint="https://your-resource.openai.azure.com",
    api_version="2024-02-01",
    proxy_key="pk_live_..."
)

Azure requires additional headers: X-Azure-Endpoint and X-Azure-Api-Version.

Proxy URL: POST /v1/proxy/azure/openai/deployments/{deployment}/chat/completions

Using the Proxy Directly

You can use Proxle without an SDK by calling the proxy endpoints directly:

curl -X POST https://api.proxle.dev/v1/proxy/openai/chat/completions \
  -H "X-Api-Key: pk_live_your-key" \
  -H "X-Provider-Key: sk-your-openai-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Required headers:

Header	Description
X-Api-Key	Your Proxle API key
X-Provider-Key	Your provider API key (pass-through, never stored)
Content-Type	application/json
X-Metadata	Optional: JSON-encoded metadata object

For Azure, also include:

Header	Description
X-Azure-Endpoint	Your Azure resource endpoint
X-Azure-Api-Version	Azure API version