MindProxyOne key, every major model

OpenAI / Claude / Gemini compatible APIs with unified auth, billing, usage and troubleshooting. Live in 5 minutes.

Up and running in 5 minutes

Register → create an API key → recharge → make your first call. With the OpenAI SDK you only change one line, the base_url.

Quick start

Cherry Studio, ChatBox, NextChat, Cursor, Open WebUI — just fill in the base URL and key.

Client guides

OpenAI-compatible (/v1/chat/completions), Claude native (/v1/messages), Gemini native (generateContent).

API reference

Image generation is billed per image; video, Midjourney and async image go through the task API — submit, then poll.

Task API

USDT recharge, balance holds, settlement on real usage, cache hits billed at the cache price. Every charge is traceable.

Billing

Every response carries a request id. Together with usage, the error-code table and rate-limit headers, issues are locatable.

Error codes