Home

Export2AI Wiki 📦

Export2AI is a VS Code and Cursor extension for AI-ready project zip archives, ChatGPT project uploads, Claude code review context, Cursor workspace handoff, and offline LLM token estimates.

The README stays compact on purpose. This wiki is the deeper human guide: visual, searchable, and organized by what a reader wants to do next.

---

Pick Your Path

I want to...	Go here
Install the extension and create my first AI context zip	🚀 Quick Start
Understand every command in the Export2AI menu	🛠️ Features & Workflow
Know what is included, excluded, replaced, or protected	🛡️ Safe Exports & Excludes
Understand .git handling and repository-control files	🌿 Git Metadata Soft-Delete
Pick a target model and interpret token counts	🧮 Token Estimates & AI Model Support
Change extension settings safely	⚙️ Settings & Configuration
Reduce context size without breaking review quality	🧹 Comment Stripping & Compression
Fix settings, zips, tokens, or file-collection issues	🧯 Troubleshooting
Contribute code or review architecture	👩‍💻 Developer Guide
Build, test, package, and release the VSIX	🧪 Build, Test & Release
Give an AI coding agent the repo safely	🤖 Agent Chokepoints

---

What Export2AI Solves

AI coding tools need a useful project snapshot. Manual zip files often include the wrong files: build folders, dependency caches, old exports, local secrets, or missing CI/test/docs context.

Export2AI creates a source-focused archive for:

Workflow	Why it helps
ChatGPT codebase analysis	Upload one clean project zip instead of pasting files manually.
Claude code review	Keep docs, tests, CI workflows, and architecture notes together.
Cursor AI workspace handoff	Give an agent a compact project archive without local machine noise.
LLM context-window planning	See an offline token estimate before you upload.
Repository validation by AI agents	Keep .github/, .gitignore, tests, docs, and AGENTS.md visible.

---

Current Documentation Map

Layer	Location	Purpose
Compact overview	README.md	short public landing page with direct wiki links
Human guide	this GitHub Wiki	searchable user docs and feature explanations
Technical docs	docs/	maintainer and contributor reference
Agent rules	AGENTS.md	constraints for AI coding agents
Tests	tests/ and scripts/test-*.js	smoke checks and regression coverage

---

Core Concepts in One Screen

Concept	Plain meaning
AI-ready zip archive	A .zip written to the workspace root with source text and useful context.
Safe defaults	Built-in ignore patterns skip common noise and credential-like material.
Soft-delete Git metadata	Real repo-control files stay; local .git/ internals become a harmless marker.
Offline token estimate	Token counting happens locally; no provider API call is made.
Target model	export2ai.llmModel controls zip names, tokenizers, menus, manifest text, and status bar labels.
Manifest	_EXPORT2AI_MANIFEST.txt explains what was counted, omitted, and redacted.

---

Good Defaults

{
  "export2ai.llmModel": "gpt-5.5",
  "export2ai.enableTokenCounting": true,
  "export2ai.softDeleteGitMetadata": true,
  "export2ai.useBuiltInExcludePatterns": true,
  "export2ai.removeComments": false,
  "export2ai.compressCode": false,
  "export2ai.showExplorerTokenBadges": false
}

These defaults preserve source readability, prevent common accidental leaks, and avoid cluttering the VS Code/Cursor Explorer with token badges unless the user opts in.

---

Fast Links

• Open VSX listing
• GitHub releases
• Root README
• Technical docs index
• License