Retio-ai/pagemap
by Various
🐍 🏠 - Compresses ~100K-token HTML into 2-5K-token structured maps while preserving every actionable element. AI agents can read and interact with any web page at 97% fewer tokens
MCP
Retio-ai/pagemap
Added 1 June 2026
Overview
Retio-ai/pagemap compresses large HTML documents (up to ~100K tokens) into compact structured maps of 2-5K tokens. It preserves all actionable elements such as links, forms, and buttons, enabling AI agents to read and interact with web pages using 97% fewer tokens.
Best for
Best for
Developers building cost-sensitive AI agents that need to parse and act on web page content
Use cases
- Reduce token costs when feeding web page content to AI agents
- Extract structured, actionable summaries of HTML for automation
- Enable AI agents to navigate and interact with web pages via compact representations
Notes
Retio-ai/pagemap compresses large HTML documents (up to ~100K tokens) into compact structured maps of 2-5K tokens. It preserves all actionable elements such as links, forms, and buttons, enabling AI agents to read and interact with web pages using 97% fewer tokens.
32 stars on GitHub. Last updated 2026-05-21.
Use cases
- Reduce token costs when feeding web page content to AI agents
- Extract structured, actionable summaries of HTML for automation
- Enable AI agents to navigate and interact with web pages via compact representations
Pros
- Reduces HTML token count by 97% while retaining critical content
- Preserves every actionable element for agent interaction
- Open-source Python tool, easy to integrate into existing pipelines
Cons
- Small user base (32 stars) suggests limited community support
- Requires pre-fetched HTML; does not handle dynamic content or browser automation
- Documentation and examples may be sparse given early stage
Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.
Pros
- Reduces HTML token count by 97% while retaining critical content
- Preserves every actionable element for agent interaction
- Open-source Python tool, easy to integrate into existing pipelines
Cons
- Small user base (32 stars) suggests limited community support
- Requires pre-fetched HTML; does not handle dynamic content or browser automation
- Documentation and examples may be sparse given early stage
Pairs with
Other entries in the index that connect to this one. Click through to see the chain.
Claude Computer Use
Anthropic
Claude takes the mouse and keyboard. A vision-based agent that controls a real desktop, not just a browser.
Claude Code
Anthropic
Anthropic's terminal-native coding agent. Reads your repo, edits files, runs tests, ships PRs.
agent-infra/mcp-server-browser
Various
The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra
browsermcp/mcp
Various
Browser MCP is a Model Context Provider (MCP) server that allows AI applications to control your browser
microsoft/markitdown
Various
Python tool for converting files and office documents to Markdown.