Stop wiring model APIs by hand.
One SDK to chain, route, and monitor GPT-4, Claude, Gemini, and any agent. With built-in retries and observability.
npm install @relayplane/sdk
import RelayPlane from '@relayplane/sdk';
// That's it! Zero configuration required
const result = await RelayPlane.ask("Explain quantum computing");
console.log(result.response.body);
console.log(result.reasoning.rationale); // See the AI's decision-making100K free calls/month • No credit card required • Bring your own keys (BYOK)
🪄 Zero-Config Wins in 30 Seconds
Experience the magic that's making developers switch from complex AI libraries
Just Ask Anything
RelayPlane.ask("Code review this function")Auto-detects your API keys, picks the best model, explains why. No configuration needed.
Ready-to-Use Patterns
RelayPlane.summarize(doc, {length: 'brief'})8+ built-in examples: chatbot, translate, codeReview, research. Copy, paste, ship.
Helpful Error Messages
❌ API key issue
💡 export OPENAI_API_KEY="sk-..."No more cryptic errors. Get copy-paste solutions and helpful recovery suggestions.
Before RelayPlane vs. After RelayPlane
See why developers are switching from complex AI libraries
Before RelayPlane
// Complex setup required
import OpenAI from 'openai';
import Anthropic from '@anthropic-ai/sdk';
const openai = new OpenAI({
apiKey: process.env.OPENAI_API_KEY
});
const anthropic = new Anthropic({
apiKey: process.env.ANTHROPIC_API_KEY
});
// Manual model selection
let response;
try {
response = await openai.chat.completions.create({
model: 'gpt-4', // Hard-coded choice
messages: [...]
});
} catch (error) {
// Manual fallback logic
if (error.status === 429) {
response = await anthropic.messages.create({
model: 'claude-3-sonnet-20240229',
messages: [...]
});
}
}❌ Complex setup and configuration
❌ Manual model selection guessing
❌ Custom fallback logic required
❌ Cryptic error messages
❌ No cost optimization
With RelayPlane v0.2
// Zero config, intelligent optimization
import RelayPlane from '@relayplane/sdk';
// That's it! Everything else is automatic
const result = await RelayPlane.ask(
"Analyze this data and suggest improvements"
);
console.log(result.response.body);
console.log(result.reasoning.rationale);
// "Selected Claude 3.5 Sonnet because:
// 40% faster for analysis tasks, $0.02 vs $0.08 cost"
// Ready-to-use patterns
const review = await RelayPlane.codeReview(myCode);
const summary = await RelayPlane.summarize(doc);✅ Zero configuration required
✅ Intelligent model selection with reasoning
✅ Automatic fallback and optimization
✅ Helpful error messages with solutions
✅ Built-in cost and latency optimization
Experience the Magic in 30 Seconds
See why developers are calling this "magical" - try our interactive CLI demo
Fast setup • Works with any AI provider • See the magic instantly
What Developers Are Saying
"Finally! An AI library that just works. The zero-config experience is magical - I was building in under 30 seconds."
"The built-in examples saved me hours. RelayPlane.codeReview() works perfectly out of the box."
"Switched from LangChain. The error messages actually help instead of confusing me more."
Why Developers Choose Relay
Universal Relay SDK
One SDK to connect Claude, GPT-4, Gemini, and more — with support for agent chaining, retries, and fallbacks.
Open-Source MCP Router
Drop-in Rust router for multi-agent chaining. MIT-licensed on GitHub and designed to be forked.
Python A2A Adapter
Seamlessly integrate Relay with LangChain and AutoGen using our pip-installable adapter.
Real-Time Logs & Cost Tracking
View latency, usage, and cost per call in your hosted dashboard. Automatically log retry paths and fallbacks.
Relay Optimize™
Route based on latency or cost. Automatically retry and fallback when models fail. Reduce costs with intelligent deduplication.
Relay Optimize™: Built-In Reliability
Production-grade failover and smart routing, built into every call
Automatically fallback
if Claude, GPT-4, or Gemini fails
Route by cost or latency
Smart routing based on your priorities
Response caching (cuts latency & spend)
with intelligent caching
await relay.optimize({
strategy: "fallback",
candidates: ["claude", "gpt-4", "gemini"],
payload: {
messages: [
{ role: "user", content: "Summarize this meeting" }
]
}
})Start Free. Scale When Ready.
Zero config, zero cost to start. Upgrade for optimization and team features.
Developer
Perfect for getting started with AI development
- 100,000 API calls per month
- 3 provider keys
- Basic relay routing
- Community support
Solo-Pro
Enhanced performance for professional developers
- 1,000,000 API calls per month
- 5 provider keys
- Relay Optimize™ & caching
- Email support
Team
Built for teams collaborating on AI applications
- 5,000,000 API calls per month
- 10 provider keys
- Team workspace
- Priority support
Need higher volumes? We have Growth and Enterprise plans with unlimited calls.
Run Your First Chain in the Playground
// Example: Chain two agents together
const summary = await relay({
agent: "summarize",
input: "Long document text..."
});
const translation = await relay({
agent: "translate",
input: summary.content,
target_language: "spanish"
});
console.log(translation.content);No setup. No keys. Just future-proof patterns for agent orchestration.
Frequently Asked Questions
What is Relay?
RelayPlane is the AI control-plane: one API for model calls, agent chains, cost/latency insights, and hosted billing. Sign up free to get started.
When can I use it?
RelayPlane is live and ready to use. Sign up free to start building with the SDK.
Do I bring my own API keys?
Yes — keep full control and pay your provider directly.
Is RelayPlane SOC 2 compliant?
SOC 2 Type I audit is in progress (ETA Q3 2025).
What do I get if I upgrade?
Unlock Relay Optimize™: retry logic, smart fallback, and latency-based routing. Add your API key — that's it.
Latest from the Blog
Insights, tutorials, and updates on AI infrastructure
Stay in the Loop
Get weekly insights on AI infrastructure, agent patterns, and RelayPlane updates. Join 500+ developers building the future of AI orchestration.
We'll send a thank-you email immediately. No spam, ever.
Building with Industry Leaders
Anthropic
OpenAI
LangChain
Vercel
Ready to experience the zero-config magic?
Join thousands of developers building with AI in under 30 seconds. No configuration, no complexity, no barriers.
Zero configuration • Get started in 30 seconds • 10,000 free calls/month