Anybrowse
Fetch any public URL as clean Markdown in the agent so research, docs, and competitive reads stay copy-pasteable for prompts.
Overview
AnyBrowse is a MCP server for the Idea phase that turns URLs into clean LLM-ready Markdown via real Chrome through remote streamable-http.
What is this MCP server?
- Remote streamable-http MCP at anybrowse.dev/mcp (v1.0.0)
- Converts arbitrary URLs to clean Markdown for LLMs
- Uses real Chrome browsers for faithful page rendering
- Reduces manual copy-paste and HTML cleanup in agent workflows
- Open-source repo at github.com/kc23go/anybrowse
- Remote endpoint: https://anybrowse.dev/mcp
- Version 1.0.0 with streamable-http transport
- Rendering: real Chrome browsers per description
Community signal: 3 GitHub stars.
What problem does it solve?
Raw HTML and broken copy-paste from websites pollute agent context and slow competitive and docs research.
Who is it for?
Solo builders who constantly feed web pages into agents for research, docs, or light competitive SEO reads.
Skip if: Bulk scraping at scale, authenticated sessions behind logins, or compliance-sensitive archival without your own legal review.
What do I get? / Deliverables
After you add AnyBrowse, your agent can request Markdown snapshots of URLs for summarization, outlining, and integration into your repo docs.
- Clean Markdown captures of requested URLs in agent chat
- Faster research summaries without HTML noise
- Reusable web excerpts for docs and content drafts
Recommended MCP Servers
Journey fit
Canonical shelf is Idea / research because AnyBrowse’s core job is turning messy web pages into LLM-ready text for discovery and analysis. Research subphase matches URL-to-Markdown reading; secondary placement covers docs authoring and SEO content harvesting.
How it compares
Chrome-backed URL-to-Markdown reader MCP, not a Swiss weather data connector or B2B sales brief generator.
Common Questions / FAQ
Who is AnyBrowse for?
Indie developers and content-minded builders who want agents to read public web pages as clean Markdown without custom browser automation.
When should I use AnyBrowse?
Use it during research, docs drafting, or SEO competitor reads whenever a URL should become structured text inside your agent thread.
How do I add AnyBrowse to my agent?
Register remote streamable-http https://anybrowse.dev/mcp as io.github.kc23go/anybrowse in your MCP configuration.