Skip to content

ARIA Gateway Documentation

Welcome to the ARIA Gateway documentation. ARIA Gateway is a high-performance, OpenAI-compatible API gateway for Bravo Zero's AI services.

Overview

ARIA Gateway provides:

  • OpenAI-Compatible API - Drop-in replacement for OpenAI API with full SDK compatibility
  • Unified Access - Single endpoint for multiple AI backends (AI-Oracle, Carousel, Athena)
  • Cost Management - Per-API-key billing, budgets, and usage tracking
  • Enterprise Features - Rate limiting, circuit breakers, request coalescing
  • Multi-Region - Global deployment with automatic failover

Quick Start

from openai import OpenAI

client = OpenAI(
    api_key="your_aria_api_key",
    base_url="https://api.bravozero.ai/v1"
)

response = client.chat.completions.create(
    model="aria-default",
    messages=[
        {"role": "user", "content": "Hello, ARIA!"}
    ]
)

print(response.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "your_aria_api_key",
  baseURL: "https://api.bravozero.ai/v1"
});

const response = await client.chat.completions.create({
  model: "aria-default",
  messages: [
    { role: "user", content: "Hello, ARIA!" }
  ]
});

console.log(response.choices[0].message.content);

curl https://api.bravozero.ai/v1/chat/completions \
  -H "Authorization: Bearer your_aria_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "aria-default",
    "messages": [
      {"role": "user", "content": "Hello, ARIA!"}
    ]
  }'

Features

OpenAI Compatibility

ARIA Gateway is fully compatible with OpenAI's API specification:

Endpoint Status
/v1/chat/completions ✅ Supported
/v1/embeddings ✅ Supported
/v1/models ✅ Supported
/v1/completions (legacy) ⚠️ Deprecated

Custom ARIA Endpoints

Beyond OpenAI compatibility, ARIA provides advanced orchestration:

Endpoint Description
/v1/aria/orchestrate Multi-model orchestration
/v1/aria/agents Agent-based workflows
/v1/aria/tools Tool/function execution
/v1/aria/sessions Stateful sessions

Architecture

graph TB
    subgraph clients [Clients]
        SDK[SDK/Client]
        App[Application]
    end

    subgraph gateway [ARIA Gateway]
        LB[Load Balancer]
        GW[API Gateway]
        Auth[Authentication]
        RL[Rate Limiter]
        CB[Circuit Breaker]
        Cache[Response Cache]
    end

    subgraph backends [AI Backends]
        Oracle[AI-Oracle]
        Carousel[Carousel]
        Athena[Athena]
    end

    SDK --> LB
    App --> LB
    LB --> GW
    GW --> Auth
    Auth --> RL
    RL --> CB
    CB --> Cache
    Cache --> Oracle
    Cache --> Carousel
    Cache --> Athena

Getting Help

What's Next?