Show HN: Optimal model routing directly in Claude, Codex and Cursor (github.com)

Weave has introduced an innovative drop-in proxy that optimally routes requests to AI models, such as Claude, Codex, and Gemini, enhancing the efficiency of accessing various AI services. By utilizing a compact on-box embedder, the router intelligently selects the best model for each request using a scoring mechanism based on the Avengers-Pro 2 algorithm. This solution aims to streamline AI integration by allowing developers to easily connect to multiple providers via a single endpoint, with support for both commercial and open-source models. This development is significant for the AI/ML community because it simplifies accessing diverse AI models while maintaining security through local key storage and encrypted data. The router, compatible with various APIs, facilitates versatile AI interactions, enhancing development workflows. For developers, integrating this router requires minimal setup, only needing to point their AI applications to a local server. With built-in observability features and adaptable configuration for different project scopes, this tool promises not only to improve model accessibility but also to aid in monitoring and managing AI performance efficiently.