model-orchestrator
Routes coding-agent tasks to cheaper or stronger models using editable rules, subagents, and a CLI runner.
Tested on Linux
Use it
Linux x86_64
You need first
- Runtimes:
node>=18
Global npm install of model-orchestrator
npm install -g model-orchestrator
Check that it works
aunx --helpThe output should contain “aunx route [--dir PATH]”.
First thing to try
npx model-orchestrator --listRun aunx route "<your task>" to get a routing suggestion for a task.
What it does
- Route routine work to cheaper modelsTested
How to use it
- Run
aunx route "<your task>"with a routine task such as a rename. - Read the suggestion line for the lane, model tier and effort level.
- Run
- Get a suggested lane for a taskTested
How to use it
- Run
aunx route "<your task>". - Check the output for the suggested lane, model tier and effort.
- Run
- Hand off scoped work to worker lanesNot shown working
- See your routing split and subagent activityNot shown working
How far pickfit got
Linux x86_64
- Installs
- Starts
- 2 of 4 features work
- Works offline
macOS: the project says it runs there; pickfit hasn’t tried it.
Good for
- route routine coding work to cheaper models
- delegate scoped work to another AI CLI lane via brief
- get a model routing suggestion for a task
- summarize local routing split and subagent activity
- run acceptance checks against delegated work
An alternative to LiteLLM, Portkey, OpenRouter and claude-code-router.
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
Its main function works
- Replayed from scratch
- Works offline
- Checked on 30 Oct 2026
- Version 1.1.2
Differences the test machine hit
- README: aunx cli-run --doctor
In the test: Exits 10 listing 6 enabled lanes with binaries MISSING when no project lanes are installed; still the correct honest doctor output.