save-md — Claude Code skill: save web pages as Markdown
save-md saves web pages, YouTube videos, Google Docs and PDFs to a .md file with source metadata. Extracts full text through direct export endpoints, bypassing summaries.
MIT 367 installs ★ 216 updated 2026-11-10
npx skills add mblode/agent-skills --skill save-md -a claude-codeWhat save-md does
The skill fetches the named source and writes it to a .md file in your project tree. Handles web pages, YouTube, Google Docs, Sheets, Slides, PDFs, DOCX and tweets. Frontmatter includes title, URL and timestamp.
A table of tested endpoints for different hosts sits inside: each comes with a ready curl line and a description of the errors it returns. For YouTube there is a workaround via --cookies-from-browser when the IP is blocked. An empty .md next to a large PDF means a scan: the skill renders pages and reads them.
Without FFmpeg or Playwright for PDF rendering the skill stops at an empty file. WebFetch cuts long texts down to 120 words — the skill bypasses it by requesting Content-Type: text/markdown directly.
What it helps with
- save an article or video to the project in one command
- fetch full text of a web page, skipping brief summaries
- extract content from Google Docs or a locked PDF
- get a YouTube video transcript with timestamps
- record the source with URL and date in frontmatter
Example prompts for the agent
- Save this article to
references/article.md - Download the video transcript and put it in
notes/ - Convert this Google Doc to Markdown
How to install save-md
1Skills CLI (Vercel)
npx skills add mblode/agent-skills --skill save-md -a claude-code2Manually: put the files in place
~/.claude/skills/save-md/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/save-md/
1Skills CLI (Vercel)
npx skills add mblode/agent-skills --skill save-md -a codex2Manually: put the files in place
.agents/skills/save-md/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add mblode/agent-skills --skill save-md -a cursor2Manually: put the files in place
.cursor/skills/save-md/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install mblode/agent-skills skills/save-md2Manually: put the files in place
.github/skills/save-md/SKILL.md1Skills CLI (Vercel)
npx skills add mblode/agent-skills --skill save-md -a gemini-cli2Platform CLI
gemini skills install https://github.com/mblode/agent-skills --path skills/save-mdManually: put the files in place
.devin/skills/save-md/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add mblode/agent-skills --skill save-md -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about save-md
How to install save-md in Claude Code?
Run npx skills add mblode/agent-skills --skill save-md -a claude-code. The skill appears in ~/.claude/skills/save-md/SKILL.md.
What does save-md do?
Saves web pages, YouTube videos, Google Docs and PDFs to a .md file with source metadata, extracting full text through direct export endpoints.
What is needed for PDFs and YouTube?
Rendering PDF scans requires Playwright. For YouTube when the IP is blocked you need --cookies-from-browser from a local machine.
Source text
SKILL.md · MIT
--- name: save-md description: Saves a named source to Markdown with provenance and faithful extraction through direct export endpoints. Use when asked to "save this article", "get the markdown", "transcribe this", or "keep this source". A URL supplied as task context alone does not trigger conversion; a chat summary stays in chat. --- # Save as Markdown Write the named source to a `.md` file, in full, and say where it is. The body comes from the fetched bytes, never from a summary. - **IS:** one named source (URL, attachment, paste, YouTube video, Google Doc, PDF) to one `.md` in the user's tree. - **IS NOT:** answering from a fetch, summarizing into chat, crawling beyond the URLs named, or reconstructing a talk from memory. ## Done A `.md` exists where the user can open it, with this frontmatter, and reading it back returns the source's last paragraph: ```yaml --- title: "<title>" source: "<URL or path>" date: "<date -u +%Y-%m-%dT%H:%M:%SZ, run now>" type: web | youtube | gdoc | sheet | slides | pdf | docx | tweet | conversation --- ``` Fetch with `curl -sSL -o <file> <url>` and read the file. Prefer a text endpoint over HTML chrome. A missing file with a named reason (paywall, login wall, bot check, private doc) beats a plausible one. | File | Read when | |------|-----------| | `references/source-endpoints.md` | Always, for the URL's host: the tested endpoint, its curl line, the failure each one shows, and the opt-in fallback reader | ## Gotchas - WebFetch hands the page to a small model and returns its answer: a 1,200-word post came back as 120 words. Only `Content-Type: text/markdown` under 100K characters passes through intact. - oEmbed, GitHub `?raw=true`, and Drive downloads redirect across hosts. Without `-L` you save an empty 301. - YouTube's "Sign in to confirm you're not a bot" is an IP block: `--cookies-from-browser` on a laptop, stop in a sandbox. The description is not the talk. - An empty `.md` next to a large PDF means a scan, not a blank document: render the pages and read them. Maintenance only: `evals/evals.json` holds scenarios for anyone changing this skill.
© mblode · MIT · mblode/agent-skills