Route every OpenClaw request to the right model with ClawRouter
ClawRouter is the open-source model router built for OpenClaw: install the plugin, fund one wallet, run /model blockrun/auto and every request is routed and paid in USDC.
How it works
ClawRouter sits between OpenClaw and every model provider. It classifies each prompt, picks the cheapest model that can handle it, compresses the prompt, caches repeated answers and pays for the call over x402 from a single wallet. OpenClaw sees one provider; the catalog behind it is every model BlockRun serves.
It is MIT-licensed and the routing decisions are auditable — the router can print which model a prompt would go to without spending. The plugin is the shortest path: install, fund, and switch the conversation's model to the auto route.
Behind it is the BlockRun catalog: 78 chat models plus image, video, music and speech generation, search, market data and multi-chain RPC, priced per call. 6 open-weight models are free and need no wallet.
Setup
1. Install as an OpenClaw plugin
openclaw plugins install @blockrun/clawrouter2. Or standalone
npm install -g @blockrun/clawrouter3. In any OpenClaw conversation
/model blockrun/autoQuestions
- Does ClawRouter change my prompts?
- It compresses them before sending and caches repeated answers; both are configurable, and the eco and quality routes trade those off differently.
- Can I pin a model instead of auto?
- Yes — /model with any catalog id pins it. Auto is the route that picks per prompt.