Setup & Model Access
Parton must be configured before it can perform AI work.
The setup gate
If you run a command that needs AI and Parton is not configured yet, it blocks execution and starts setup.
You can also run setup directly:
parton setup
Supported model access modes
Parton supports two access modes:
CLI access (recommended)
This is the preferred path for most local users.
Why:
- you may already be logged into a provider CLI
- you may already pay for that tool
- it avoids extra API setup friction
API access
You can also configure direct API access with a provider key.
This is useful when:
- you prefer direct API usage
- you want a specific stage to use a specific provider
- you do not want to rely on a local provider CLI
Dynamic model discovery
Parton does not rely on hardcoded model names.
During setup it discovers models dynamically from the selected provider.
That means:
- available models are fetched live
- the setup TUI shows real choices
- you select from a list instead of guessing a model string
A manual fallback may exist, but the main UX is selection from discovered models.
Per-stage model selection
Parton can use different models for different stages:
- Planning
- Execution
- Judge
Example strategy:
- Planning → strong reasoning model
- Execution → coding-focused model
- Judge → cheaper review model
Doctor
After setup, run:
parton doctor
Doctor verifies:
- provider access
- selected model availability
- command reachability for CLI-based providers
- API key validity for API-based providers
- validation command readiness
- writable local state directories
Config behavior
Parton supports:
- project-local config
- global fallback config
Project config overrides global defaults.
That lets you:
- keep a default provider setup globally
- override model choices per repo when needed
Example first run
parton run "create todo app"
If setup is missing, Parton will guide you through:
- provider selection
- model selection
- validation
- saving config
Then it continues into clarification and planning.
Common setup issues
Provider CLI not detected
Make sure the CLI tool is installed and available in PATH.
Model list does not load
This usually means:
- provider access is invalid
- session is expired
- the provider command failed
- API access is missing or invalid
API key invalid
Re-enter the key and run:
parton doctor
Setup completed but run still blocks
This usually means the config was not saved correctly or the selected model failed validation.
Run:
parton doctor