# Weave Router

> ELv2 on-box embedder proxy that routes each prompt to the right model

**Category:** AI Agents  
**Pricing:** source-available  
**URL:** https://router.workweave.ai  
**GitHub:** https://github.com/weave-os/router  
**Added:** 2026-09-20

---

## The Hook

Weave Router is a drop-in Anthropic/OpenAI/Gemini proxy that picks a model per request with a tiny on-box embedder instead of a vibes-based router prompt.

## What's Good

Go proxy from Weave that sits on the endpoint you already use, claims sub-50ms routing, and targets cost cuts by sending easy work to cheaper models. Useful when an agent farm pays for one default frontier model on every call.

## The Catch

Elastic License 2.0 is source-available, not OSI open-source — you may not offer Weave Router as a hosted/managed service under ELv2 without permission. Vs LiteLLM: LiteLLM is the broader multi-provider gateway/SDK; Weave Router's wedge is per-prompt model selection via on-box embeddings. Routing quality still depends on your traffic mix and evaluation.

## Verdict

ELv2 source-available model router. Strong cost lever for agent fleets; no managed offering without a license review.