InfiniLoom Documentation

InfiniLoom is a high-performance repository context generator for large language models. Transform your codebase into optimized formats for Claude, GPT-4, Gemini, Llama, Mistral, DeepSeek, Qwen, and other AI models.

Installation

# Using Cargo (Rust)
cargo install infiniloom

# Using Homebrew
brew tap Topos-Labs/infiniloom
brew install --cask infiniloom

# Using pip (Python)
pip install infiniloom

# Using npm (Node.js)
npm install -g infiniloom

Quick Start

# Pack a repository for Claude
infiniloom pack ./my-project --model claude --output context.xml

# Generate a repository map
infiniloom map ./my-project

# Scan for secrets
infiniloom scan ./my-project --security-check

Commands

Output Formats

InfiniLoom supports multiple output formats optimized for different use cases:

• XML — Claude-optimized structured format
• Markdown — GPT-optimized human-readable format
• JSON — Machine-readable structured data
• YAML — Gemini-optimized long context format
• TOON — Most token-efficient compressed format
• Plain — Simple text output

Supported Models

Token counting and optimization for:

• Anthropic Claude (Claude 3, Claude 4)
• OpenAI GPT (GPT-4, GPT-4o)
• Google Gemini
• Meta Llama
• Mistral
• DeepSeek
• Qwen
• Cohere
• Grok