Aether is a self-hosted AI API gateway that provides teams and individuals with multi-tenant management, intelligent Load Balancer, cost quota control and health monitoring capabilities. Through a unified API entry, you can seamlessly connect with mainstream AI services such as Claude, OpenAI, and Gemini and their CLI tools. Q&A Q: How to turn on/off request body records? The administrator configures the level of detail of logging in the system settings: level recorded content Base Basic request information Headers Base + request header Full Headers + Request Body Q: How do administrators configure 1M context/ 1H caching capability support for models? Model management: Set up 1M context/ 1H cache capability support for models and configure prices Provider management: Add a key to the endpoint that supports this capability and check the corresponding capability label Q: How do users use 1H cache? Model level: Enable 1H caching policy for specified models in model management Key level: Use 1H cache policy for specified keys in key management Note: If a key is set to force 1H caching, the key can only be used with a model that supports 1H caching. Matching the provider Key will result in this Key not being used for Claude Code, Codex, and GeminiCLI at the same time, because it is more recommended to use the model to enable 1H caching. Q: How to configure Load Balancer? Switch scheduling modes in managing backend provider management: Q: Will the provider's free plan billing model include costs? No. The charging model of the free package has a multiplier of 0, and the generated records are not included in the cost.
