FeaturesHow It WorksPricingDocsInstallContact UsGitHub

Parton CLI is in early alpha

Things may break or change between releases. We don't recommend using Parton on production projects yet. If you run into anything unexpected, we'd genuinely appreciate your feedback — it helps us make Parton better for everyone.

Troubleshooting

This page covers the most common problems when using Parton.

Install problems

parton: command not found

The install directory may not be in your PATH.

A common fix:

export PATH="$HOME/.local/bin:$PATH"

Then restart your shell and try again:

parton version

Permission denied during install

The installer may not have permission to write to the target directory.

Try again and follow the permission prompt, or use a user-writable install path.

Unsupported OS or architecture

Parton currently supports:

  • macOS Apple Silicon
  • macOS Intel
  • Linux x64

Setup problems

Provider CLI not detected

Make sure the provider CLI (e.g. claude, codex) is installed and available in your shell PATH.

Model list does not load

Possible causes:

  • expired provider login
  • invalid provider CLI session
  • missing API key
  • provider outage

Run setup again:

parton setup

API key invalid

Reconfigure setup or export the correct key:

parton setup

env_key collision

If you use multiple API-based providers, make sure each has a unique env_key in parton.toml to avoid environment variable collisions.

Run problems

Clarification keeps asking questions

Your prompt may be too ambiguous.

Try being more specific. Instead of:

parton run "create app"

say:

parton run "create a Next.js TypeScript todo app with local state and no auth"

Planning failed

Check:

  • setup is complete (parton setup)
  • provider is reachable
  • clarification completed
  • the prompt is specific enough

Scaffold failed

If scaffold files fail to generate:

  • check provider connectivity
  • check that the model is available
  • try --verbose for debug output

Auto-fix loop not converging

If validation keeps failing after auto-fix retries:

  • inspect the error output — the validation command may have a pre-existing issue
  • check that validation_commands in the plan are correct
  • try a more capable model for the execution stage

Provider problems

CLI provider not responding

Make sure the provider CLI is installed, logged in, and reachable:

  • claude — requires Anthropic CLI login
  • codex — requires OpenAI Codex CLI login

API provider errors

For OpenAI API or other API-based providers:

  • verify the API key is valid
  • check that the env_key in parton.toml matches your exported environment variable
  • check provider status pages for outages

Ollama not connecting

Make sure Ollama is running locally:

ollama list

If the model is not pulled yet:

ollama pull llama3

Validation problems

Validation failed after auto-fix

This means Parton tried to auto-fix validation failures but could not converge within the retry limit.

Check:

  • the validation error output
  • whether the validation commands themselves are correct
  • whether the task is too complex for the selected model

Dependencies not installed

If validation fails because of missing packages, check that the plan's install_command ran successfully. You can run it manually:

npm install
# or whatever the project uses

Debugging workflow

  1. Run with --verbose for detailed output
  2. Check the CLI output for error messages
  3. Verify setup with parton setup
  4. Try a simpler prompt to isolate the issue
  5. Check provider connectivity independently

Related