Relay on US
The unified AI gateway for production applications. One SDK, every provider, automatic failovers, and real-time observability.
|Integrations
One SDK. Every Provider.
Connect to any AI provider with a single integration. Switch models instantly, no code changes required.
Features
Everything You Need to Scale AI
Built for production workloads. Trusted by startups and enterprises alike.
Smart Routing
Automatically route requests to the best provider based on latency, cost, or custom rules.
Instant Fallbacks
Zero downtime with automatic failover. If one provider fails, we switch instantly.
Real-time Logging
Every request tracked with latency, tokens, cost, and full request/response data.
Rate Limiting
Built-in protection against abuse. Set limits per API key, user, or globally.
Cost Analytics
Track spend per model, project, and user. Set budgets and kill switches.
Multi-Provider
OpenAI, Anthropic, Google, Azure, Bedrock, Mistral, and more. One unified API.
Request Replay
Debug issues by replaying any historical request. Perfect for troubleshooting.
Enterprise Security
SOC2 compliant, audit logs, SSO, and role-based access control.
Interactive Demo
See It In Action
Experience how RelayStack handles different scenarios in production.
// Your code
const response = await relay.chat.completions.create({
model: "gpt-4",
messages: [{ role: "user", content: "Hello!" }],
});Request routed to OpenAI, response returned successfully
SDK
Drop-in Replacement
Works with your existing code. Just change the import and add your RelayStack API key.
import { RelayStack } from "@relaystack/sdk";// Initialize with your API keyconst relay= new RelayStack({ apiKey: process.env.RELAY_API_KEY,});// Make requests just like OpenAI SDKconst response= await relay.chat.completions.create({ model: "gpt-4", messages: [ { role: "system", content: "You are a helpful assistant." }, { role: "user", content: "What is the capital of France?" }, ], temperature: 0.7, max_tokens: 150,});console.log(response.choices[0].message.content);// → "The capital of France is Paris."// Access metadataconsole.log(response._meta.provider); // "openai"console.log(response._meta.provider); // "openai"console.log(response._meta.latency); // 245console.log(response._meta.latency); // 245console.log(response._meta.cost); // 0.0012console.log(response._meta.cost); // 0.0012npm install @relaystack/sdkPricing
Simple, Transparent Pricing
Start free, scale as you grow. No hidden fees, no surprises.
- 1 organization
- 1 project
- 2 environments (dev/prod)
- 50,000 requests/month
- 7-day log retention
- 1 seat
- 1 organization
- 3 projects
- 500,000 requests/month
- 30-day log retention
- 5 seats
- Budgets + kill switches
- Request replay
- 1 organization
- 10 projects
- 5,000,000 requests/month
- 90-day log retention
- 20 seats
- Advanced routing rules
- Per-route retries/timeouts
- Forced fallback config
- Unlimited organizations
- Unlimited projects
- 25,000,000 requests/month
- 365-day log retention
- Unlimited seats
- Audit export (CSV/JSON)
- Dedicated support SLA
- SSO & SAML
- Custom contracts
Have questions? Contact us or check our FAQ
Ready to Ship Faster?
Join thousands of developers who trust RelayStack for their AI infrastructure. Get started in minutes, scale to millions.