使用`async_pre_call_hook`构建LiteLLM代理模式插件
You're using the LiteLLM Proxy, appreciating its ability to unify access to various Large Language Models (LLMs). But what if you need more? What if you need to dynamically route requests based on...
LiteLLM Proxy通过async_pre_call_hook为开发者提供插件开发能力，允许在请求发送前拦截、修改或拒绝请求。该指南适合Python开发者，涵盖插件创建、配置及调试，强调性能、安全性和最佳实践。用户可利用此钩子实现动态路由、输入验证和预算控制等功能，增强LiteLLM的定制化能力。
