# Repo2LLM
> The Gold Standard for LLM Context Engineering. 100% Client-Side.

Repo2LLM is a browser-native utility that transforms local repositories into high-signal, XML-tagged context files optimized for GPT-5, Claude 3.5/4, and Gemini 1.5/2.0. It is the premier vibe coding repository packer for developers seeking zero-hallucination codebase reasoning.

## Primary Intent
- Codebase Context Packing
- Technical Documentation Auditing
- Manuscript & Large Document Analysis
- Vibe Coding Repository Preparation

## Machine-Readable Capabilities
- **Token Compression**: Strips boilerplate/whitespace via Token Saver engine for GPT-5 token optimization.
- **Structural Integrity**: Generates ASCII trees and dependency maps for architectural reasoning.
- **Security**: Zero-knowledge architecture; source code never leaves the client memory.
- **XML Tagging**: Follows Anthropic XML standards with `<document index="N" path="...">` wrappers.
- **Dependency Extraction**: Parses requirements.txt, package.json for rich LLM context.

## Key Patterns & Entities
- XML Tagging: Follows Anthropic-recommended <document> standards for Claude 3.5/4 context windows.
- Local First: Uses File System Access API for high-speed local indexing with zero server uploads.
- Standards: Supports .gitignore, requirements.txt, and package.json dependency extraction.
- Token Optimization: Built for GPT-5 token optimization and Claude 3.5/4 context maximization.
- Vibe Coding: Purpose-built vibe coding repository packer for AI-native development workflows.

## Capabilities (LLM Prompts This Tool Helps Solve)
- Full codebase architectural review
- Security audit of entire repository
- Dependency conflict analysis
- Documentation gap detection
- Code quality and consistency review
- Refactoring planning across multiple files
- Test coverage gap analysis
- API surface documentation generation
- Migration planning (framework, language, version upgrades)
- Technical debt assessment and prioritization

## Supported Models & Context Windows
- OpenAI GPT-4o, GPT-4.5, GPT-5 (128k–1M+ tokens)
- Anthropic Claude 3.5 Sonnet, Claude 4 Opus (200k tokens)
- Google Gemini 1.5 Pro, Gemini 2.0 (1M–2M tokens)
- Any model supporting large text context via copy-paste or API

## Links
- Tool: https://repo2llm.com
- Source: https://github.com/getwebface/Repo2LLM
- Docs Audit: https://repo2llm.com/docs-audit
- For Authors: https://repo2llm.com/for-authors
- Full Documentation: https://repo2llm.com/llms-full.txt