Gateway Engine v1.0.0 Pro · Active

Developer Documentation

Unified routing specifications, endpoint protocols, and code integration guides for high-scale LLM operations.

Connect to Anthropic, OpenAI, and Google Gemini models through a resilient, multi-account routing gateway engineered for zero-latency streaming pass-through and automated failover.

ROUTER_DAEMON // v1.0.0 Pro
MULTI-CLOUD
CLIENT INGRESSEDGE GATEWAY1,420 req/s · SSECLAUDEANTHROPIC14ms TTFT · 99.98%OPENAIMODELS18ms TTFT · 99.99%GEMINIGOOGLE11ms TTFT · 100.0%EZYBUSROUTER CORE
Throughput1,420 tok/s
Latency12 ms
Upstream3/3 Healthy
FailoverAuto
Unified API Base URLProduction Ready
https://api.ezybus.org/api/v1

Core Architectural Advantages

Zero-Latency Pass-Through

SSE chunk bytes are flushed directly to the client socket without buffering to prevent Cloudflare 524 timeouts.

Automated 429 Failover

Instant fallback across multi-tier accounts when rate limits or upstream 503 provider incidents are detected.

Thinking Token Telemetry

Accurate downstream tracking and token attribution for Claude extended thinking and reasoning loops.

Anthropic Prompt Caching

Full transparency for prompt caching write and read headers, dramatically slashing compute latency and costs.

Quickstart Integration

gateway-client.tsOpenAI SDK / Cursor IDE
import OpenAI from 'openai';

// Point standard OpenAI SDK to EzyBus unified gateway
const openai = new OpenAI({
  baseURL: 'https://api.ezybus.org/api/v1',
  apiKey: process.env.EZYBUS_API_KEY || 'sk-eb-your-key-here',
});

const stream = await openai.chat.completions.create({
  model: 'gpt-5.6',
  messages: [{ role: 'user', content: 'Explain distributed consensus in 2 sentences.' }],
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content || '');
}

Live Endpoint Mappings

POST/api/v1/chat/completions

OpenAI Chat Completions endpoint. Compatible with OpenAI SDK, Cursor, VSCode extensions, and custom tooling.

Required Headers
Authorization: Bearer <API_KEY>Content-Type: application/json
POST/api/v1/messages

Anthropic Messages API. Seamlessly routes Claude prompts, cache control blocks, and extended thinking budgets.

Required Headers
x-api-key: <API_KEY>anthropic-version: 2023-06-01Content-Type: application/json
POST/api/gemini/v1beta/models/{model}:streamGenerateContent

Native Google Gemini content generation endpoint. Pass through multimodal requests and system instructions directly.

Required Headers
x-goog-api-key: <API_KEY>Content-Type: application/json
GET/api/v1/models

Discovers active model pools, supported capabilities, context window lengths, and token pricing rates.

Required Headers
Authorization: Bearer <API_KEY>

Latency Benchmarks

Realtime
Connection Pooling Boost+42% throughput
Median Time-To-First-Token< 380ms

Gateway Integrity

uTLS Fingerprint EmulationActive
AES-256 GCM Key ShardingEnforced
Zero-Copy SSE Socket FlushingEnabled
Thinking Token Parser & MeteringActive
Loki Realtime Audit TelemetryStreaming

Need Integration Support?

Running behind a custom reverse proxy or Nginx? Ensure you pass X-Accel-Buffering: no to preserve realtime SSE streaming.

Console Dashboard