Stop 403 Forbidden Errors

Your AI agent deserves better than a 404.

Stop letting websites block your tools. md-fetch uses your real browser to bypass anti-scraping and deliver perfectly cleaned Markdown to your LLMs.

Standard Scraper
$ curl https://blocked-site.com
HTTP/1.1 403 Forbidden
<!-- Cloudflare protection enabled -->
Agent: "I'm sorry, I cannot access that website due to security restrictions."
Recommended
md-fetch (Chrome mode)
$ md-fetch --browser chrome https://blocked-site.com
# Main Article Title
This content was fetched using a real headless browser session, bypassing blocks...
Agent: "I've analyzed the site. Here is the summary you requested..."

Built for the AI Era

Traditional scrapers are dead. md-fetch is designed to bridge the gap between the complex web and LLM context windows.

01. Real Browsers

Bypass Everything

Don't just spoof headers. Use real Headless Chrome or Firefox to execute JavaScript, handle cookies, and bypass sophisticated bot detection.

02. Token Efficiency

LLM-Optimized Output

HTML is 90% noise. md-fetch strips scripts, ads, and CSS, delivering clean Markdown that maximizes your agent's context window.

03. Developer First

CLI + REST API

Use it as a one-off CLI tool or run md-fetch serve to deploy a high-performance concurrent processing API for your production pipelines.

Get Started in Seconds

No complex configuration. Just works.

1

One-liner (Linux & macOS)

curl -fsSL https://raw.githubusercontent.com/nathabonfim59/md-fetch/main/install.sh | sh
2

Using Go

go install github.com/nathabonfim59/md-fetch@latest
3

For AI Agents (OpenCode / Gemini CLI / Claude Code)

Add md-fetch as a specialized skill to your agent ecosystem:

npx skills add nathabonfim59/md-fetch

Ready to unblock your agents?

Star on GitHub