-
Notifications
You must be signed in to change notification settings - Fork 131
Description
Description
I want to be able to pick the provider and model when running in non-interactive mode
Use Case
If you want to use nanocoder with multiple instances of the same CLI tool, you need to be able to pass provider and model details to the run command.
https://getpolyscope.com/
https://github.com/aarondfrancis/counselors
Here are two tools that you need CLI flags for.
It would also be nice to start the nanocoder session with a specific provider or model selected.
Proposed Solution
Just allow the model and provider to be set when using run
nanocoder --provider openrouter --model google/gemini-3.1-flash-lite-preview run "my prompt goes here still"If my default provider + model are local, this allows me to quickly switch to a remote one when necessary without having to run through different slash commands or editing the config or ENV.
Alternatives Considered
I saw some issues for allowing the ENV to control config. I think this works nicely with both of those options as well.
Additional Context
- I have searched existing issues to ensure this is not a duplicate
- This feature aligns with the project's goals (local-first AI assistance)