Human Use MCP server enables your AI agents to leverage real-time human intelligence through the Rapidata API. It exposes a set of human-in-the-loop tools that let you solicit authentic human feedback for AI tasks, improving decision quality and alignment. Available tasks include short free-text responses to questions, image classification, and image ranking, as well as comparison of two texts to select the better option. Each tool engages real people to perform the task, providing reliable ground truth for evaluation and guidance. The MCP is designed to be integrated into your workflow via Cursor using the human-use server, by running the Rapidata script rapidata_human_api.py with a specified directory. The project ships with a custom Streamlit app to interact with the MCP server, addressing compatibility issues seen with other clients. A hosted version is also available at chat.rapidata.ai, enabling you to access features without setting up your own environment. Setup involves cloning the repo, configuring environment variables, creating a virtual environment, installing dependencies with uv, and launching the app.