Submit a tool →
W

Weave Router

by Weave (WeaveOS)

Routes each coding-agent request to the cheapest model that can still do the job.

Model APIs & Infra Free, source-available self-host; hosted plan takes 5% of routed spend

What is Weave Router?

Weave Router sits between your coding agent (Claude Code, Codex, or Cursor) and the models it calls. It scores how hard each request actually is, then sends it to the cheapest model likely to get it right, escalating back to a frontier model if the cheap one falls short.

Version 2.0 adds routing across multiple provider subscriptions at once, so it can run Claude models inside Codex or GPT models inside Claude Code, draining whichever flat-rate plan still has quota left before it touches a metered API key.

Key features

  • Per-turn complexity classifier decides which model handles a request
  • Cache-aware switching, only reroutes when the savings beat the cache-rebuild cost
  • Routes across multiple provider subscriptions, not just API keys
  • Automatic escalation to a frontier model if a cheaper one is failing the task
  • One command configures Claude Code, Codex, and Cursor at once
  • Source-available under the Elastic License 2.0, so it can be self-hosted

How to get started

  1. Run npx @workweave/router to detect and configure Claude Code, Codex, and Cursor
  2. Add the provider API keys or subscriptions you want it to route across
  3. Keep using your coding agent as normal; routing happens per request
  4. Check the dashboard to see how much of your traffic is being routed and saved

Use cases

  • Teams with a large monthly Claude Code, Codex, or Cursor bill looking to cut cost without changing workflow
  • Developers who want to combine multiple subscriptions instead of paying per token on top of a flat-rate plan
  • Anyone who'd rather not manually pick a model for every coding task
  • Self-hosting a router instead of depending on a third party to see your traffic

Weave Router vs. alternatives

ToolHow it compares
Weave RouterRoutes each coding-agent request to the cheapest model that can still do the job.
OpenRouterOpenRouter routes any API request across 300+ models by price or preference; Weave Router is narrower but specifically understands and routes across Claude Code/Codex/Cursor's flat-rate subscription seats, not just API keys.

FAQ

Is Weave Router free to use?

Weave Router's plans: Free, source-available self-host; hosted plan takes 5% of routed spend. Pricing and free-tier limits change over time, so check the official site above for the latest details.

What is Weave Router used for?

Routes each coding-agent request to the cheapest model that can still do the job.

Who is Weave Router best for?

Weave Router is a good fit for teams with a large monthly Claude Code, Codex, or Cursor bill looking to cut cost without changing workflow, or developers who want to combine multiple subscriptions instead of paying per token on top of a flat-rate plan.

What are the alternatives to Weave Router?

Commonly compared alternatives include OpenRouter, see the comparison above for how they differ.

Related tools

Last updated: 2026-09 · Reviewed by AIKetra editors · How we evaluate tools · Embed a Featured badge