AIProxy Relay - OpenAI-compatible API at 90% less cost

Hi everyone,

I built AIProxy Relay to solve a problem I was facing myself: API costs were eating up my side-project budget.

What it does:

- Drop-in replacement for OpenAI's API – just change the base_url

- Supports gpt-4, qwen-plus, qwen-turbo

- Costs ~$5 per 1M tokens (output) – about 1/10 of GPT-4

Why I built it:

I was spending too much on API calls while building AI apps. This relay routes requests to more affordable models while keeping the same interface.

Looking for beta testers:

I'm giving away free tokens to early users in exchange for feedback.

Interested? Comment below or DM me!