How it works
Three ideas, no dashboard to babysit.
Wendl sits between your team's chat and the model providers. Every request passes through it; every dollar is metered by it; nothing about how your team works has to change.
Right-size every request
Wendl reads each task and routes it to the smallest capable model — instead of paying top-tier prices for a job that never needed them.
Routine work stays home
Cron jobs, summaries, the everyday grind run on local models on your own hardware. Free to run, private by default, no tokens burned.
Every dollar is visible
Type /limits for your caps and spend, /stats for what you've saved versus paying full freight. Honest numbers, in the room where you work.
Managed — not set-and-forget
The best model is a moving target. We keep you on it.
A capable new open-weight model lands every few weeks, and the price-to-performance frontier shifts under you. Staying on top of it is a full-time job — so we make it ours, benchmarking each new model and harness as it ships against the way your team actually works.
- Backtested on your own traffic. Before we move you to something new, we replay your real past work against it and score cost and quality. You switch only when it wins — measurably.
- Models and harnesses, both. Not just which model, but which agent framework and which routing thresholds. We upgrade the whole stack underneath you.
- Tuned to your team, continuously. The more you run, the better the routing fits your work. You get the results; we do the tuning.
You never look back — the numbers already made the call.
Try it — free & open source
Put the numbers in your own chat.
Those /limits and /stats commands are a free, open-source OpenClaw plugin. It reads your own gateway and prints your caps, spend, and savings right in the chat — no agent call, no tokens burned. Install it straight from GitHub:
$ openclaw plugins install git:github.com/wendl-ai/openclaw-wendl@main $ openclaw plugins enable wendl
Then add your litellmMasterKey under plugins.entries.wendl.config in openclaw.json, restart the gateway, and type /limits. Full setup and the routing config are in the plugin repo. Not on OpenClaw yet? See a sample breakdown →
- /limits — caps & spend
- /stats — money saved
- /wendl — get a managed setup
Ready when you are
Self-host it free, or let us run it.
The gateway is a single Node process — no Docker, no Postgres. Start with your own numbers, upgrade to managed when you want the frontier watched for you.
Get started → or email [email protected]