Control every key
Set what each key can use, how much, and for how long.
- Model allowlist
- Spending limit
- Expiry date
Leading model vendors
- Anthropic
- OpenAI
- DeepSeek
- xAI
- Zhipu AI
Leave the hard parts to the gateway
Your app talks to one address. The gateway handles the rest.
Automatic failover
Moves to the next channel the moment one fails.
Both protocols
Native OpenAI and Anthropic support.
One line to change
Swap the base_url and you are connected.
See every call
Model, tokens and cost, recorded per request.
- Filter by key and model
- 30-day usage trend
- Per-request detail
Pay for what you use
Usage-based billing with a clear balance history.
- Online top-up
- BTC / USDT top-up
- Redemption codes
Models and pricing
Input, output and cached tokens are priced separately.
Prices in ¥ per 1M tokensSee all pricing
Works with the tools you already use
14 clients supported, each with a setup guide.
Connect in three steps
