Integrations / SDK & Client Integrations

OpenAI-Compatible Apps

Point OpenAI- and Anthropic-compatible SDKs and applications at Aurora's /v1 endpoints.

Browse docs

Overview

Most applications that support an OpenAI- or Anthropic-compatible base URL can use Aurora by changing only the base URL, API key, and model name.

Aurora exposes:

/v1/chat/completions
/v1/responses
/v1/embeddings
/v1/rerank
/v1/files for providers that support files
/v1/batches for providers that support batches
/v1/models

Base URL

Use your Aurora origin plus /v1 as the SDK base URL when the SDK expects a versioned OpenAI base path:

text

http://your-aurora-host/v1

Use a managed API key or master key as the OpenAI API key value.

Chat example

text

curl http://your-aurora-host/v1/chat/completions \
  -H "Authorization: Bearer $AURORA_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-aurora-User-Path: /team/alpha" \
  -d '{
    "model": "regular",
    "messages": [{"role": "user", "content": "Reply with ok"}]
  }'

Embeddings example

text

curl http://your-aurora-host/v1/embeddings \
  -H "Authorization: Bearer $AURORA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": ["Hello from Aurora"]
  }'

Rerank example

text

curl http://your-aurora-host/v1/rerank \
  -H "Authorization: Bearer $AURORA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "rerank-model",
    "query": "best gateway feature",
    "documents": ["model aliases", "unrelated text"],
    "top_n": 1,
    "return_documents": true
  }'

Use aliases

Prefer aliases such as regular, reasoning, or embedding-default in client configuration. Then operators can move aliases between providers without application redeploys.

OpenAI SDK setup

Point the official OpenAI Python or Node.js SDK at Aurora by setting base_url (or baseURL):

Python

text

from openai import OpenAI
 
client = OpenAI(
    base_url="http://your-aurora-host/v1",
    api_key="sk-aurora-...",
)
 
response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Node.js

text

import OpenAI from "openai";
 
const client = new OpenAI({
  baseURL: "http://your-aurora-host/v1",
  apiKey: "sk-aurora-...",
});
 
const response = await client.chat.completions.create({
  model: "gpt-4o-mini",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(response.choices[0].message.content);

What works

Aurora supports the OpenAI chat surface end-to-end:

Chat Completions (/v1/chat/completions)
Embeddings (/v1/embeddings)
Models (/v1/models)
Streaming (SSE)
Tool / function calling
Responses API (/v1/responses)

Aurora does not implement legacy POST /v1/completions text completions. Use chat completions instead, or provider passthrough at /p/openai/v1/completions if you must reach the upstream endpoint directly.

Anthropic SDK setup

If you are using the Anthropic SDK, point it at the origin (not at /v1). The SDK appends /v1/messages automatically. See Anthropic Messages API for the full setup.

← All docs

Integrations / SDK & Client Integrations

OpenAI-Compatible Apps

Point OpenAI- and Anthropic-compatible SDKs and applications at Aurora's /v1 endpoints.

Browse docs

Overview

Most applications that support an OpenAI- or Anthropic-compatible base URL can use Aurora by changing only the base URL, API key, and model name.

Aurora exposes:

/v1/chat/completions
/v1/responses
/v1/embeddings
/v1/rerank
/v1/files for providers that support files
/v1/batches for providers that support batches
/v1/models

Base URL

Use your Aurora origin plus /v1 as the SDK base URL when the SDK expects a versioned OpenAI base path:

text

http://your-aurora-host/v1

Use a managed API key or master key as the OpenAI API key value.

Chat example

text

curl http://your-aurora-host/v1/chat/completions \
  -H "Authorization: Bearer $AURORA_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-aurora-User-Path: /team/alpha" \
  -d '{
    "model": "regular",
    "messages": [{"role": "user", "content": "Reply with ok"}]
  }'

Embeddings example

text

curl http://your-aurora-host/v1/embeddings \
  -H "Authorization: Bearer $AURORA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": ["Hello from Aurora"]
  }'

Rerank example

text

curl http://your-aurora-host/v1/rerank \
  -H "Authorization: Bearer $AURORA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "rerank-model",
    "query": "best gateway feature",
    "documents": ["model aliases", "unrelated text"],
    "top_n": 1,
    "return_documents": true
  }'

Use aliases

Prefer aliases such as regular, reasoning, or embedding-default in client configuration. Then operators can move aliases between providers without application redeploys.

OpenAI SDK setup

Point the official OpenAI Python or Node.js SDK at Aurora by setting base_url (or baseURL):

Python

text

from openai import OpenAI
 
client = OpenAI(
    base_url="http://your-aurora-host/v1",
    api_key="sk-aurora-...",
)
 
response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Node.js

text

import OpenAI from "openai";
 
const client = new OpenAI({
  baseURL: "http://your-aurora-host/v1",
  apiKey: "sk-aurora-...",
});
 
const response = await client.chat.completions.create({
  model: "gpt-4o-mini",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(response.choices[0].message.content);

What works

Aurora supports the OpenAI chat surface end-to-end:

Chat Completions (/v1/chat/completions)
Embeddings (/v1/embeddings)
Models (/v1/models)
Streaming (SSE)
Tool / function calling
Responses API (/v1/responses)

Anthropic SDK setup

If you are using the Anthropic SDK, point it at the origin (not at /v1). The SDK appends /v1/messages automatically. See Anthropic Messages API for the full setup.