workweave/router
Cut AI costs 40-70% with one endpoint swap
WorkWeave Router exists to stop agentic systems from overpaying for every prompt by routing each one to the cheapest adequate model in milliseconds.
Collecting fresh signals — velocity needs a few days of history.
collecting data…
star history
What it does WorkWeave Router is a Go-based AI gateway built for agentic systems. It exposes an OpenAI-compatible endpoint so you can reroute prompts from tools like Claude Code or Codex to whichever model is cheapest or fastest for the job. The project claims sub-50 ms routing latency and 40–70% cost savings.
The interesting bit The entire sell is friction reduction: change one URL and the router handles model selection without touching client code. That’s a compelling pitch for teams running autonomous coding agents that currently default to the most expensive model for every request.
Key highlights
- Go implementation, ~1.1k stars
- Targets agentic workflows (Claude Code, Codex, Anthropic, OpenAI)
- Drop-in endpoint swap; claims no client rewrites needed
- Advertises <50 ms routing overhead
- Claims 40–70% cost reduction via automatic model selection
Verdict Give it a spin if your agentic pipeline is hemorrhaging API credits and you want a low-effort optimization. Pass if you need granular routing policies or deep observability—the public docs are too thin right now to confirm how much control you actually get.
Frequently asked
- What is workweave/router?
- WorkWeave Router exists to stop agentic systems from overpaying for every prompt by routing each one to the cheapest adequate model in milliseconds.
- Is router open source?
- Yes — workweave/router is an open-source project tracked on heatdrop.
- What language is router written in?
- workweave/router is primarily written in Go.
- How popular is router?
- workweave/router has 1.2k stars on GitHub.
- Where can I find router?
- workweave/router is on GitHub at https://github.com/workweave/router.