Autorouter CLI
The autorouter CLI lets you try LiteLLM's complexity-based auto-routing locally, against models your key already has access to on your real, running proxy. It picks a cheaper or more expensive model depending on how complex a prompt looks, without editing your proxy's config.yaml and without any request bypassing it. It stands up a throwaway local proxy that forwards every request back to your real proxy, then points Claude Code at that local proxy for the session.
The autorouter CLI is early and evolving. Tell us what works, what breaks, and what you want next in the Autorouter discussion on GitHub. Your feedback directly shapes where this goes.
1. Install the CLI​
Install lite from the internal staging branch with a single curl command; uv is bootstrapped automatically if it is missing.
curl -fsSL https://raw.githubusercontent.com/BerriAI/litellm/litellm_internal_staging/scripts/install-cli.sh | \
LITELLM_CLI_REF=litellm_internal_staging sh
2. Point the CLI at Your Proxy​
Set your real proxy URL and key. The autorouter forwards every request through this proxy with this key.
export LITELLM_PROXY_URL=http://localhost:4000
export LITELLM_PROXY_API_KEY=sk-...
3. Configure the Autorouter​
Run the interactive wizard. It discovers the model groups your key can reach and asks you to assign models to each complexity tier.
lite autoroute configure
4. Run the Autorouter​
Start the ephemeral local proxy. It runs in the foreground and streams routing decisions live, so you can watch which tier and model gets picked for each request.
lite autoroute up
5. Run Claude Code as Normal​
In another tab, start Claude Code. The autorouter now automatically intercepts all Claude Code traffic.
claude
For the full reference, including recovery from an unclean shutdown and important caveats, see the autorouter CLI README.