AST-aware code context engine for RAG, vector databases, and AI assistants. Tree-sitter parsing, PageRank ranking, BLAKE3 content-addressable chunks. Integrates with Pinecone, Weaviate, Qdrant. Supports Claude, GPT-5, GPT-4o, Gemini, Llama (27+ tokenizers
npm install infiniloomHigh-performance repository context generator for LLMs.
Transform your codebase into optimized context for Claude, GPT-4o/GPT-5, Gemini, and other Large Language Models.
``bash`
npm install -g infiniloom
`bashPack repository into Claude-optimized XML
infiniloom pack /path/to/repo --format xml
Commands
| Command | Description |
|---------|-------------|
|
pack | Transform repository into LLM-optimized context |
| scan | Analyze repository statistics |
| map | Generate PageRank-based symbol map |
| index | Build symbol index for fast diff context |
| diff | Get context for code changes |
| impact | Analyze change impact |
| chunk | Split repository into manageable pieces |
| init | Create configuration file |
| info | Show version and configuration info |Supported Models
- OpenAI: GPT-5.x, GPT-4o, O3, O1
- Anthropic: Claude
- Google: Gemini
- Meta: Llama, CodeLlama
- Others: Mistral, DeepSeek, Qwen, Cohere, Grok
Output Formats
- XML - Optimized for Claude (prompt caching)
- Markdown - Optimized for GPT models
- YAML - Optimized for Gemini
- JSON - For programmatic access
- TOON - Most token-efficient (~40% smaller)
Alternative Installation
If npm installation fails, you can install via:
`bash
Homebrew Cask (macOS - fast, pre-built binary)
brew tap Topos-Labs/infiniloom
brew install --cask infiniloomHomebrew Formula (builds from source)
brew tap Topos-Labs/infiniloom
brew install infiniloomCargo (Rust)
cargo install infiniloom
``- Documentation
- GitHub Repository
MIT