One OpenAI-compatible endpoint for Claude, GPT and Gemini: notes from building APIGOTO

I work on the APIGOTO team; APIGOTO is independent, not affiliated with OpenAI, Anthropic or Google.

Selam EthioDev! I want to share how we approached a problem many of us hit when building with LLMs: every provider has its own SDK, errors and billing.

What APIGOTO does:

• one OpenAI-compatible base_url: https://api.china-aiapi.com/v1
• Claude, GPT and Gemini behind the same API shape
• automatic failover, capped at 3 retries (we capped it so a failing request returns instead of looping)
• a log entry for every request
• BYOK (use your own provider keys) or pay-per-use

Minimal Python example: from openai import OpenAI client = OpenAI(base_url="https://api.china-aiapi.com/v1", api_key="<your key>") resp = client.chat.completions.create(model="<model>", messages=[{"role": "user", "content": "Hello"}])

I'd love feedback from developers in Addis and beyond: is 3 retries the right default? What do you need in per-request logs?

www.china-aiapi.com → https://www.china-aiapi.com/grokbot