New in v0.2: Zero-Config Wins & Zero-Config AI

RelayPlane SDK

The easiest way to add AI to your app. Zero configuration, intelligent model selection, built-in examples, and automatic optimization. Get from idea to production in minutes.

Zero-Config Win

Get started in under 30 seconds. No configuration, no setup, no complexity. Just ask and get intelligent AI responses.

Install & Use

npm install @relayplane/sdk

import RelayPlane from '@relayplane/sdk';

// That's it! Zero configuration required
const result = await RelayPlane.ask("Explain quantum computing simply");
console.log(result.response.body);
console.log(result.reasoning.rationale); // See why this model was chosen

✨ Auto-Detection

Automatically discovers your API keys and selects the best model for each task

🧠 Intelligent Routing

Analyzes your prompt and routes to the optimal model with reasoning

Ready-to-Use Examples

Common AI patterns ready to use. Copy, paste, and customize for your needs.

💬 Chatbot

const chat = await RelayPlane.examples.chatbot(
  "Hello! How can you help me?",
  {
    personality: "You are a helpful coding assistant",
    budget: "moderate"
  }
);

📄 Summarization

const summary = await RelayPlane.examples.summarize(
  longDocument,
  {
    length: "brief",
    focus: "key findings"
  }
);

🔍 Code Review

const review = await RelayPlane.examples.codeReview(
  myCode,
  {
    language: "typescript",
    focus: "security"
  }
);

🌍 Translation

const translation = await RelayPlane.examples.translate(
  "Hello world",
  "Spanish",
  { tone: "formal" }
);

8+ Examples Available: chatbot, summarize, codeReview, translate, research, creativeWriting, dataAnalysis, explain, and more!

Contextual Error Handling

When things go wrong, get helpful suggestions instead of cryptic error messages.

Example Error Output

❌ OpenAI API key not found or invalid

🔧 Quick Fix:
export OPENAI_API_KEY="your_api_key_here"

💡 Suggestions:
  • Add OPENAI_API_KEY to your environment variables
  • Or configure RelayPlane hosting with: RelayPlane.configure({ apiKey: "rp_your_key" })
  • Check that your API key is valid and has the right permissions
  • Verify OpenAI account has sufficient credits/quota

📚 Helpful Links:
  • https://docs.relayplane.com/setup/api-keys
  • https://docs.relayplane.com/quick-start

🔑 API Key Issues

Clear guidance on setting up API keys with copy-paste solutions

⚡ Rate Limits

Automatic fallback suggestions and retry guidance

Advanced Features

Intelligent Model Selection

const result = await RelayPlane.ask(
  "Review this React component for security issues",
  {
    budget: "minimal",      // Cost optimization
    priority: "quality",    // Favor accuracy
    taskType: "coding"      // Hint for better selection
  }
);

// See the reasoning
console.log(result.reasoning.rationale);
// "Selected Claude 3.5 Sonnet because: excellent for 
//  code analysis, estimated $0.003 cost, ~1500ms response time"

console.log(result.reasoning.alternatives);
// Array of alternative model suggestions with reasons

Automatic Optimization

// Configure global optimization
RelayPlane.configure({
  defaultOptimization: {
    enabled: true,
    strategy: "balanced",
    maxRetries: 3,
    enableCache: true,
    cacheTtl: 300
  }
});

// All requests now have:
// ✅ Automatic retries with fallbacks
// ✅ Intelligent caching
// ✅ Cost optimization
// ✅ Performance monitoring

Global Optimization Settings

Configure advanced optimization settings across your entire application:

import RelayPlane from '@relayplane/sdk';

const relayplane = new RelayPlane({
  apiKey: 'your-api-key',
  globalOptimize: {
    cacheStrategy: {
      type: 'hybrid',           // memory, disk, hybrid, none
      maxSize: '500MB',
      evictionPolicy: 'lru',    // lru, lfu, fifo
      compression: true,
      encryption: true
    },
    batchProcessing: {
      enabled: true,
      defaultConcurrency: 5,
      maxBatchSize: 100,
      autoGrouping: true,
      timeoutMs: 30000
    },
    parallelExecution: {
      enabled: true,
      maxConcurrentRequests: 10,
      requestPooling: true,
      loadBalancing: 'round_robin'  // round_robin, least_connections, random
    },
    timeoutSettings: {
      requestTimeoutMs: 30000,
      connectionTimeoutMs: 5000,
      readTimeoutMs: 15000,
      totalTimeoutMs: 45000,
      retryTimeoutMs: 60000
    },
    rateLimitingStrategy: {
      strategy: 'adaptive',         // fixed, adaptive, burst
      backoffMultiplier: 1.5,
      maxBackoffMs: 30000,
      jitterEnabled: true,
      respectProviderLimits: true,
      globalRateLimit: 1000         // requests per minute
    }
  }
});

Agent-to-Agent (A2A) Communication

Enable agents to communicate with each other for complex multi-step workflows:

import { relay } from '@relayplane/sdk';

// Multi-agent workflow with A2A communication
const analysisResult = await relay({
  to: 'claude-3-7-sonnet-20250219',
  payload: {
    max_tokens: 1000,
    messages: [
      { role: 'user', content: 'Analyze this sales data and identify trends' }
    ]
  },
  metadata: {
    workflow_id: 'sales-analysis-pipeline',
    next_agent: 'report-generator-v2',  // A2A routing
    context: {
      data_source: 'Q4-2024',
      format: 'executive-summary'
    }
  }
});

// Chain multiple agents
const reportResult = await relay({
  to: 'gpt-4.1',
  payload: {
    max_tokens: 2000,
    messages: [
      { 
        role: 'system', 
        content: 'You are a report generator. Use the analysis results to create an executive summary.' 
      },
      { 
        role: 'user', 
        content: `Analysis results: ${analysisResult.body.content}` 
      }
    ]
  },
  metadata: {
    workflow_id: 'sales-analysis-pipeline',
    parent_request: analysisResult.relay_id,
    final_step: true
  }
});

Available on Team+ plans: A2A communication enables sophisticated multi-agent workflows with automatic context passing and error handling.

Installation & Quick Setup

1. Install the SDK

npm install @relayplane/sdk

2. Set API Keys (Optional)

# Add any provider's API key - RelayPlane will auto-detect
export OPENAI_API_KEY="your-openai-key"
export ANTHROPIC_API_KEY="your-anthropic-key"
export GOOGLE_API_KEY="your-google-key"

# Or use RelayPlane hosting for all models
export RELAY_API_KEY="rp_your-key"

3. Start Using

import RelayPlane from '@relayplane/sdk';

// Zero-config magic
const result = await RelayPlane.ask("What is the capital of France?");

// Built-in examples
const summary = await RelayPlane.examples.summarize(document);
const chat = await RelayPlane.examples.chatbot("Hello!");

// Advanced features
const optimized = await RelayPlane.optimize(request);
const chain = await RelayPlane.chain({ steps: [...] });

API Reference

RelayPlane.ask()

Zero-config method for instant AI access with intelligent model selection.

RelayPlane.ask(input: string, options?: {
  budget?: 'unlimited' | 'moderate' | 'minimal';
  priority?: 'speed' | 'balanced' | 'quality';
  taskType?: 'general' | 'coding' | 'creative' | 'analysis' | 'translation';
  stream?: boolean;
  config?: Partial<EnhancedRelayConfig>;
}): Promise<{
  response: RelayResponse;
  reasoning: {
    selectedModel: string;
    rationale: string;
    alternatives: Array<{model: string, reason: string}>;
    detectedCapabilities: string[];
  };
}>

RelayPlane.examples

Ready-to-use patterns for common AI tasks.

Available Examples:

• chatbot(message, options)
• summarize(text, options)
• codeReview(code, options)
• translate(text, language, options)

More Examples:

• research(topic, options)
• creativeWriting(prompt, options)
• dataAnalysis(data, options)
• explain(topic, options)

Need Help?

Experiencing issues with timeouts, model errors, or slow performance? Our comprehensive troubleshooting guide covers all common issues and solutions.

🔧 View Troubleshooting Guide Fix Model Errors Optimize for Slow Machines

Migrating to v0.2

✅ Fully Backward Compatible

All existing v1.x code continues to work without changes.

// v1.x code still works
const response = await RelayPlane.relay({
      to: 'claude-3-7-sonnet-20250219',
  payload: { messages: [{ role: 'user', content: 'Hello!' }] }
});

// New v0.2 Zero-Config Wins are additive
const result = await RelayPlane.ask("Hello!"); // New!

🎯 Recommended: Try the New Features

• Replace manual model selection with RelayPlane.ask()
• Use RelayPlane.examples.* for common patterns
• Benefit from enhanced error handling automatically