↓Skip to main content
  1. Tools/
  2. Code Artifacts/

gh-md

·3 mins·
Ai-Ready Code Artifacts Git-Only Github-Wiki Github Markdown Backup Wiki
Language: Go
License: MIT
Maturity: beta
Verified: 2026-02
Table of Contents

Overview #

gh-md is a tool for backing up GitHub wiki pages and other markdown content from GitHub repositories. GitHub wikis are separate git repositories (REPO.wiki.git) that are often overlooked during backups despite containing critical project documentation, onboarding guides, and design decisions.

gh-md focuses specifically on extracting and organizing this markdown content into a clean, portable directory structure that can be version-controlled independently of the source repository.

Key Features #

  • Wiki backup – clones and organizes GitHub wiki content into structured local directories.
  • Markdown extraction – pulls markdown files from repositories, preserving directory structure.
  • Clean output – produces plain markdown files that are portable and tool-agnostic.
  • Batch processing – can process multiple repositories or organizations.

Basic Usage #

# Clone a GitHub wiki directly (wikis are separate git repos)
git clone https://github.com/ORG/REPO.wiki.git

# Or use gh-md for structured backup
gh-md backup --repo owner/repo --output ./wiki-archive/

Alternative Approaches #

GitHub wikis can also be archived through several other methods:

  • Direct git clone – every GitHub wiki is a git repository at https://github.com/OWNER/REPO.wiki.git. A simple git clone captures the full wiki with history.
  • python-github-backup – includes --wikis flag for wiki backup as part of a comprehensive repository backup.
  • GitHub API – the REST API provides access to wiki page content, though not as conveniently as a direct clone.

For most archival scenarios, the direct git clone of the .wiki.git repository is the simplest and most complete approach. gh-md adds value when you need structured extraction or batch processing across many repositories.

git-annex / DataLad Integration #

Integration level: git-only.

Wiki and markdown content is small text that belongs in git proper, not in git-annex. The content is inherently diffable and benefits from line-level version tracking.

To archive wikis into a DataLad dataset:

# Create a dataset for documentation archives
datalad create docs-archive
cd docs-archive

# Clone the wiki as a subdirectory
git clone https://github.com/ORG/REPO.wiki.git wikis/REPO/

# Or use gh-md
gh-md backup --repo owner/repo --output ./wikis/REPO/

# Save with DataLad
datalad save -m "Archive wiki for ORG/REPO" wikis/REPO/

For periodic updates:

datalad run -m "Update wiki archive for ORG/REPO" \
    --output "wikis/REPO/" \
    "cd wikis/REPO && git pull origin master"

AI Readiness #

Level: ai-ready.

Markdown is one of the most AI-friendly formats:

  • Direct consumption – LLMs can read and understand markdown natively. No parsing, conversion, or preprocessing is needed.
  • Rich structure – headings, lists, code blocks, and links provide semantic structure that helps LLMs understand document organization.
  • Knowledge extraction – wiki pages often contain institutional knowledge (setup guides, architecture decisions, troubleshooting tips) that is extremely valuable for AI-assisted project understanding.
  • Search and retrieval – plain text markdown integrates easily with vector databases and RAG (retrieval-augmented generation) pipelines.

Archived wikis are ideal candidates for building project-specific knowledge bases that AI assistants can reference during development.