When using uvx, eyelet runs in an isolated environment. Here's how to ensure you're using the latest version:
# Always fetches the latest version from PyPI
uvx eyelet@latest [command]uvx eyelet --version- Persistent installation with uv tool:
# Install eyelet as a uv tool
uv tool install eyelet@latest
# Force reinstall to update
uv tool install --reinstall eyelet@latest
# Then run commands
eyelet [command]- Specify exact version:
uvx eyelet@0.3.5 # Replace with desired version- Clear cache and run:
# Clear the uv cache
uv cache clean
# Run with latest (will re-download)
uvx eyelet@latest [command]uvxis an alias foruv tool run- it runs tools in isolated environments- Tools are cached locally in
~/.local/share/uv/tools/ - Each invocation checks for the specified version
- Using
@latestensures you get the newest PyPI release - There is no
--forceflag foruvx- just use@latestto get updates
If you're not getting the latest version:
- Clear uv cache:
uv cache clean- Check PyPI for latest version:
# See available versions
uv pip index eyelet- Use uv tool for persistent installation:
# Install/reinstall as a tool
uv tool install --reinstall eyelet@latestIf you installed with pipx instead:
# Update to latest
pipx upgrade eyelet
# Or reinstall
pipx reinstall eyeletIf you're running from source with mise:
# Update dependencies
mise run update
# Or manually with uv
uv syncEyelet includes automatic update checking (can be disabled in settings):
# Check if updates are available
eyelet version --check-updates
# Disable update checks
eyelet config set auto_update_check false- Pin versions for CI/CD:
uvx eyelet@0.3.4 # Specific version for reproducibility- Use latest for interactive work:
uvx eyelet@latest # Always get newest features- Check changelog before major updates:
# View changelog online
eyelet docs --changeloguvxis essentiallyuv tool run- it runs tools in isolated environments- Tools are cached, so
@latestmight use cached version unless forced - The
--forceflag ensures a fresh installation - Each tool gets its own virtual environment with locked dependencies
- Updates are intentionally explicit to avoid breaking changes
For projects using mise, you can configure eyelet as a tool:
# mise.toml or .mise.toml
[tools]
python = "3.11"
"pipx:eyelet" = "latest" # Or specific version like "0.3.4"Then use:
mise install # Install/update all tools
mise run eyelet -- [command] # Run eyelet