Infiniloom

Transform codebases into optimized context for LLMs

Installation · Quick Start · Features · Documentation

---

Infiniloom extracts code, symbols, and structure from repositories and outputs them in formats optimized for Claude, GPT, Gemini, and other LLMs.

• Fast — Pure Rust, processes 1000+ files in <500ms
• Smart — AST-based symbol extraction (21 languages), PageRank ranking
• Secure — Automatic secret detection and redaction
• Flexible — XML, Markdown, JSON, YAML output formats

Installation

npm install -g infiniloom      # Cross-platform (recommended)
cargo install infiniloom       # Rust users
brew tap Topos-Labs/infiniloom && brew install --cask infiniloom  # macOS

Method	Command	Notes
npm	npm install -g infiniloom	Easiest, cross-platform
Homebrew Cask	brew install --cask infiniloom	macOS, pre-built binary
Homebrew Formula	brew install infiniloom	macOS, builds from source
Cargo	cargo install infiniloom	Rust users
pip	pip install infiniloom	Python library
npm library	npm install infiniloom-node	Node.js library

From source:

git clone https://github.com/Topos-Labs/infiniloom.git
cd infiniloom && cargo build --release

Quick Start

# Pack repository for Claude (XML format)
infiniloom pack . --format xml --output context.xml

# Scan repository statistics
infiniloom scan .

# Generate symbol map with PageRank ranking
infiniloom map . --budget 2000

# Get context for staged changes (AI code review)
infiniloom diff . --staged --include-diff

Output Formats

infiniloom pack . --format xml       # Claude (prompt caching hints)
infiniloom pack . --format markdown  # GPT-4/GPT-4o
infiniloom pack . --format yaml      # Gemini
infiniloom pack . --format json      # Programmatic use
infiniloom pack . --format toon      # Token-efficient (~40% smaller)

Performance

Repository	Files	Tokens	Time
infiniloom	174	667K	420ms
medium project	1,200	1.8M	1.8s

Benchmarks on M2 MacBook Pro, including full AST parsing and symbol extraction.

Features

Model-Specific Optimization

Model	Format	Optimizations
Claude	XML	Prompt caching hints, CDATA sections
GPT-4/4o	Markdown	Tables, code fences, hierarchical headers
Gemini	YAML	Query at end, structured hierarchy
Any	TOON	~40% smaller than JSON

Smart Filtering

infiniloom pack . --include "*.rs" --exclude "tests/*"
infiniloom pack . --max-tokens 50000      # Token budget
infiniloom pack . --top-files 50          # Limit to N most important files
infiniloom pack . --compression aggressive # Remove comments, docstrings

Security Scanning

Automatically detects and redacts API keys, tokens, private keys, and credentials:

infiniloom pack . --security-check   # Scan and report
infiniloom pack . --redact-secrets   # Redact with [REDACTED]

Git Integration

infiniloom pack . --include-logs --logs-count 10  # Recent commits
infiniloom pack . --include-diffs                  # Uncommitted changes
infiniloom pack github:facebook/react              # Remote repos

Diff Context for AI Code Reviews

infiniloom index .                    # Build symbol index
infiniloom diff . --staged            # Context for staged changes
infiniloom diff . HEAD~1              # Context for last commit
infiniloom diff . main..feature       # Context for branch diff
infiniloom impact . src/auth.rs       # What depends on this file?

Language	Symbols Extracted
Python	Functions, Classes, Methods, Decorators
JavaScript/TypeScript	Functions, Classes, Interfaces, Types
Rust	Functions, Structs, Enums, Traits, Impl blocks
Go	Functions, Methods, Structs, Interfaces
Java	Classes, Interfaces, Methods, Enums
C/C++	Functions, Structs, Classes
C#	Classes, Methods, Interfaces, Properties
Ruby	Classes, Modules, Methods
PHP	Classes, Functions, Methods
Kotlin	Classes, Functions, Interfaces
Swift	Classes, Structs, Functions, Protocols
Scala	Classes, Objects, Traits, Functions
Haskell	Functions, Types, Data, Typeclasses
Elixir	Modules, Functions, Macros
Clojure	Functions, Macros, Vars
OCaml	Functions, Modules, Types
Lua	Functions, Tables
R	Functions, Classes
Bash	Functions

Exact tokenization (tiktoken):

• GPT-5.2, GPT-5.1, GPT-5, O4-mini, O3, O1 (o200k_base)
• GPT-4o, GPT-4o-mini (o200k_base)
• GPT-4, GPT-3.5-turbo (cl100k_base)

Calibrated estimation:

• Claude (Anthropic)
• Gemini (Google)
• Llama 3/4, CodeLlama (Meta)
• Mistral, Mixtral (Mistral AI)
• DeepSeek V3/R1
• Qwen (Alibaba)
• Cohere Command R+
• Grok (xAI)

Level	Reduction	What's Removed
none	0%	Nothing
minimal	10-20%	Empty lines, trailing whitespace
balanced	30-40%	Comments, redundant whitespace
aggressive	50-60%	Docstrings, inline comments
extreme	70-80%	Everything except signatures

Language Bindings

Python

import infiniloom

context = infiniloom.pack("/path/to/repo", format="xml", model="claude")
stats = infiniloom.scan("/path/to/repo")

Node.js

const { pack, scan } = require('infiniloom-node');

const context = pack('./repo', { format: 'xml', model: 'claude' });
const stats = scan('./repo');

Configuration

Create .infiniloom.yaml with infiniloom init:

output:
  format: xml
  model: claude
  compression: balanced
  token_budget: 100000

scan:
  include: ["*.rs", "*.py", "*.ts"]
  exclude: ["tests/*", "docs/*"]

security:
  scan_secrets: true
  redact_secrets: true

See Configuration Guide for all options.

Documentation

Document	Description
Command Reference	All CLI commands
Configuration	Config files and environment variables
Output Formats	Format specifications
Architecture	System design

Contributing

Contributions welcome! See CONTRIBUTING.md.

cargo test --workspace    # Run tests (800+)
cargo clippy --workspace  # Lint
cargo fmt --all           # Format

License

MIT — see LICENSE.

Acknowledgments

• Tree-sitter — Fast, reliable parsing
• tiktoken-rs — Accurate token counting
• Aider — Repo-map concept inspiration

---

Made with care by Topos Labs