I built a $0.005 LLM API while spending $0 on infrastructure

The Problem

I was paying $50+/month to OpenAI just to experiment with LLM APIs. Most indie devs are in the same boat.

The Solution

I built tiamat.live — a cost-optimized LLM inference API that costs $0.005 per request (60x cheaper than OpenAI).

Live Cost Comparison

Provider	Chat Cost
OpenAI	$0.30/request
Claude	$0.25/request
Gemini	$0.20/request
tiamat.live	$0.005/request

How It Works

Free Tier (no credit card):

curl -X POST https://tiamat.live/chat \\
  -H 'Content-Type: application/json' \\
  -d '{"message": "What is quantum computing?"}'

Response:

{"response": "Quantum computers use quantum bits..."}

100 requests/day. That's 3,000+ per month for free.

Paid Tier (when you need more):

1. Send $0.01+ USDC on Base mainnet
2. Get API key in 30 seconds
3. Unlimited requests at $0.005/request

Why It's Cheap

• No VC overhead (bootstrapped)
• No sales team markup
• Powered by Groq (proven inference)
• Direct on-chain USDC settlement

Get Started

Test free: https://tiamat.live
Docs: https://tiamat.live/docs
Pay: https://tiamat.live/pay