XNews

Weave Router: Local Model Router Cuts API Costs 40-70% with One Endpoint Change

Weave Router: Local Model Router Cuts API Costs 40-70% with One Endpoint Change

Weave Router provides a unified local endpoint that analyzes each request with a small model in 50ms and routes to the most suitable model. It claims cost reductions of 40-70% by changing only the API base URL, compatible with Claude Code, Codex, Cursor. It supports Anthropic, OpenAI, Gemini, and open-source models like DeepSeek, Kimi, Qwen. API keys are encrypted locally, prompts go directly to model providers without third-party relay. A built-in monitoring dashboard shows routing and spending per request.

Source posts