Skip to content

Declarative Module Management

Instead of installing modules one at a time, manage all your project's AI context modules declaratively using a .lola-req file. This works like requirements.txt for pip or package.json for npm.

Create a .lola-req

Create a .lola-req in your project root with one module per line:

# .lola-req - AI context modules for this project

# Module names from registry or marketplace
python-tools>=1.0.0
git-workflow

# Target specific assistants
web-scraper>>claude-code
code-review>>cursor

# Direct git URLs
https://github.com/user/custom-module.git
git+https://github.com/user/another-module.git

# Git URLs with branch/tag references
https://github.com/user/module.git@main
git+https://github.com/user/[email protected]

Sync Modules

# Install all modules from .lola-req
lola sync

# Dry-run to see what would be installed
lola sync --dry-run

The sync command is idempotent - running it multiple times produces the same result.

Version Constraints

  • ==1.0.0 - Exact version
  • >=1.0.0 - Greater than or equal
  • ~1.2.0 - Compatible with 1.2.x (>= 1.2.0, < 1.3.0)
  • ^1.2.0 - Compatible with 1.x.x (>= 1.2.0, < 2.0.0)

Assistant Targeting

Use >> to target specific assistants:

module-name>>claude-code    # Install only to Claude Code
module-name>>cursor         # Install only to Cursor
module-name                 # Install to all detected assistants