I was spending $200/month on OpenAI API for my side projects. Switched to Chinese models (DeepSeek, Qwen, GLM) and my bill dropped to $15.
Problem: managing 5 different API keys and billing systems was painful.
Solution: built ChinaLLM — a single OpenAI-compatible endpoint that aggregates 20+ Chinese LLMs.
- GLM-4-Flash: $0.01/M tokens
- DeepSeek V4 Pro: $0.85/M
- Qwen-Turbo: $0.08/M
- Pay with USDT or PayPal, $5 minimum
Running on a $20/month VPS in Hong Kong. Built on new-api (33k stars).
Would love feedback from the community!