BYOK: How to Scale AI Agents Without Limits

The Scaling Problem

When building AI agents for production, you eventually hit limits:

• Rate limits from API providers
• Monthly request quotas
• Expensive token pricing
• Vendor lock-in
• Lack of control

These constraints prevent you from scaling when your agent succeeds.

The BYOK Solution

BYOK (Bring Your Own Key) flips the model:

Your Infrastructure

You provide:

• Groq API key for LLM inference
• Tavily API key for web search
• Your own rate limits and quotas

We provide:

• Intelligent routing engine
• Schema validation
• Webhook delivery
• Monitoring and logging

Benefits

Unlimited Scaling

Scale based on your provider limits, not ours:

// No rate limits from us
const response = await deepResearch({
  query: "Analyze 1000 companies",
  // Limited only by your Groq/Tavily quotas
})

Zero Markup

Pay provider pricing directly:

• Groq: $0.59/1M tokens
• Tavily: $5/1000 searches
• No markup from us

Full Control

Choose your models:

• Llama 3.1 8b (fast, cheap)
• Llama 3.3 70b (high quality)
• Mix and match as needed

Integration Guide

Step 1: Get API Keys

# Get Groq key
https://console.groq.com/

# Get Tavily key
https://tavily.com/

Step 2: Configure Deep Research

// In your agent code
const DEEP_RESEARCH_API_KEY = process.env.DEEP_RESEARCH_API_KEY
const GROQ_API_KEY = process.env.GROQ_API_KEY
const TAVILY_API_KEY = process.env.TAVILY_API_KEY

Step 3: Use the API

const response = await fetch('/api/v1/deep-research', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${DEEP_RESEARCH_API_KEY}`,
    'X-Groq-Key': GROQ_API_KEY,
    'X-Tavily-Key': TAVILY_API_KEY
  },
  body: JSON.stringify({
    query: "Your research query",
    schema: { /* your schema */ }
  })
})

Pricing Comparison

*Subject to your provider's fair use policy

Get Started

Take control of your AI agent infrastructure.

Get Your API Key

Read BYOK Documentation