Charlotte

Charlotte is an MCP (Model Context Protocol) server that enables AI agents to access and interact with the web in a structured, token-efficient manner. Instead of sending the full accessibility tree for every page, Charlotte provides compact, structured summaries, allowing agents to request only the specific details they need. This results in greatly reduced token and definition overhead, especially on large, content-heavy pages. Charlotte is ideal for developers and organizations building or integrating autonomous AI agents that need efficient, programmatic web access.