Time to First Request: 5 Minutes

1Deploy the Sovereign Stack (2-Minute Quickstart)

Selixes is containerized for zero-friction local development. Clone the repository and boot the full sovereign stack (API Gateway, Redis cache, Postgres telemetry, and Web Console) with a single command:

git clone https://github.com/selixes/gateway.git
cd gateway && docker compose up -d

The API Gateway is immediately live at http://localhost:4000 and the companion dashboard console at http://localhost:3000.

22-Line SDK Drop-in

Selixes is fully compatible with the standard OpenAI SDK. Swap the `baseURL` and point your existing application to the local reliability layer:

import OpenAI from 'openai';

// Selixes integration takes exactly 2 lines:
const openai = new OpenAI({
  apiKey: process.env.SELIXES_API_KEY || 'sk_selixes_local',
  baseURL: 'http://localhost:4000/v1'      // Swapped base URL
});

3Dispatch Your First Reliable Request

Pass transaction cost caps, timeout limits, and semantic caching directives directly through standard HTTP request headers:

const response = await openai.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Extract details from CSV invoices' }],
  headers: {
    'x-selixes-session-id': 'session_crm_batch_02',
    'x-selixes-max-session-cost': '0.15', // Autonomous block if spend exceeds $0.15
    'x-selixes-semantic-cache': 'true',   // Serve from vector DB if similar prompt exists
    'x-selixes-timeout': '5000'           // Instantly failover to Anthropic if OpenAI takes > 5s
  }
});

4Access Live Telemetry & Observability

Selixes doesn't just proxy requests; it tells you *why* routing decisions, circuit-breaker trips, and failovers occurred.

Open http://localhost:3000 in your browser to inspect active provider health, review transaction traces, and audit token economics.

curl http://localhost:4000/health