API Gateway Docs
Welcome to the EzyBus platform developer documentation. Connect to OpenAI, Anthropic, and Gemini models using a single, unified high-concurrency API gateway optimized for low-latency streaming.
Core Advantages
Zero-Latency Pass-Through
Response stream bytes are flushed directly to the socket to prevent proxy latency and Cloudflare 524 timeouts.
Automated Failover
Transparent routing automatically swaps upstream keys when hitting 429 Rate Limits or 503 Provider downtime.
Thinking & Reasoning Tokens
Native parsing and cost attribution of Claude 3.7 Sonnet thinking tokens for accurate downstream billing.
Context-Aware Caching
Fully integrates Anthropic prompt caching headers to cut system-prompt latency and compute overhead.
Quickstart Integration
import OpenAI from 'openai';
const openai = new OpenAI({
baseURL: 'https://api.ezybus.org/v1',
apiKey: 'sk-eb-your-key-here', // Your EzyBus API Key
});
const completion = await openai.chat.completions.create({
model: 'gpt-4o',
messages: [{ role: 'user', content: 'Hello!' }],
stream: true,
});Live Endpoint Mappings
OpenAI Chat Completions endpoint. Integrates fully with OpenAI SDKs, Cursor, and VSCode plugins.
Authorization: Bearer <API_KEY>Content-Type: application/jsonAnthropic Messages endpoint. Translates and forwards payloads seamlessly to Claude models with cache settings.
x-api-key: <API_KEY>anthropic-version: 2023-06-01Lists all upstream models currently configured and active in EzyBus routing pools.
Authorization: Bearer <API_KEY>Latency Benchmarks
Gateway Integrity
Need Help?
Encountering connectivity issues or 524 timeouts? Make sure to set X-Accel-Buffering: no if hosting behind custom reverse proxies.