llmifier • #

A Dart CLI tool to extract, structure, and consolidate Dart & Flutter project files into a single, LLM-friendly context.

Key Features #

• 🎯 Targeted Extraction: Focuses on Dart/Flutter projects, understanding typical structures (lib/, bin/, test/, etc.).
• ⚙️ Flexible Filtering: Uses glob patterns (like .gitignore) for precise control over included and excluded files.
• 🧠 Semantic Ordering: Prioritizes files logically (Docs, Metadata, API, etc.) for improved LLM comprehension (customizable via config).
• 🏷️ Per-File Versioning: Embeds the pubspec.yaml version in each file marker (<!-- BEGIN FILE [vX.Y.Z]: path/to/file -->) for unambiguous context.
• ✨ Extraction Modes:
  • full: Extracts the complete content of matched files.
  • api: Extracts only the public API surface (classes, functions, variables) along with documentation comments, powered by the analyzer package. Ideal for providing usage context.
• 🔧 Configuration: Uses a clear llmifierrc.yaml file alongside CLI arguments (CLI overrides YAML, which overrides defaults).
• 🚀 Init Command: Quickly generates a default llmifierrc.yaml configuration file (llmifier --init).

Installation #

Global Activation for CLI usage #

dart pub global activate llmifier

Now you can run llmifier from anywhere.

Usage #

# Generate output in the current directory (default: llms.txt)
llmifier

# Specify output file
llmifier -o my_project_context.txt

# Specify project directory
llmifier -p path/to/your/project

# Use API extraction mode
llmifier -m api

# Use API mode with a specific output file
llmifier --mode=api --output=llms-api.txt

# Get verbose logging during processing
llmifier -l

# Generate a default config file in the current directory
llmifier -i

# Show help message
llmifier -h

Why the [v1.2.3] tag in markers? This small detail is vital! It ensures the LLM always knows exactly which version of your code a specific file belongs to, preventing confusion when analyzing changes or comparing different versions side-by-side.

Configuration (llmifierrc.yaml) #

Create a llmifierrc.yaml file in your project root (or generate one using llmifier --init) to customize behavior.

Configuration Hierarchy: CLI Arguments > llmifierrc.yaml > Default Settings

Use Cases #

• LLM Code Generation: Provide comprehensive project context for generating new features or fixing bugs.
• LLM Code Review: Give the LLM a structured overview for more accurate analysis and suggestions.
• LLM-Powered Documentation: Generate technical documentation or summaries based on the actual codebase structure.
• Onboarding: Help new team members (or an LLM assistant) quickly grasp the project layout and key components.
• Version Comparison: Feed outputs from two different versions to an LLM for detailed change analysis.

Contributing #

Contributions are welcome! Please feel free to open an issue or submit a pull request on the GitHub repository.

Reporting Issues #

Please report any bugs or feature requests on the GitHub issue tracker.

License #

This project is licensed under the MIT License - see the LICENSE file for details.