Human-use

MCP Server# Overview

The MCP server is a tool that allows you to connect your AI agents with human intelligence via the Rapidata API.

Tools

1. get_free_text_responses - Will ask actual humans to provide some short free text responses to the question.
2. get_human_image_classification - Will ask actual humans to classify the images in the directory.
3. get_human_image_ranking - Will ask actual humans to rank the images in the directory.
4. get_human_text_comparison - Will ask actual humans to compare two texts and select which one is better.

Configuration

Cursor add the following to your cursor mcp.json file (usually in ~/.cursor/mcp.json)

{
  "mcpServers": {
    "human-use": {
      "command": "uv",
      "args": ["--directory", "YOUR_ABSOLUTE_PATH_HERE", "run", "rapidata_human_api.py"]
    }
  }
}

You should now be able to see the human-use server in Cursor settings.

App## Overview

The app is a custom Streamlit app that allows you to use the MCP server. We have built because of issues with other clients. Namely the Claude desktop app.

App Setup Clone Repository

git clone https://github.com/RapidataAI/human-use.git

Environment Configuration

Copy the .env.example file to .env and fill it in with your own credentials/settings

Note: paths should be ABSOLUTE paths

Installation with UV# Prerequisites

Install uv if you haven't already:

curl -LsSf https://astral.sh/uv/install.sh | sh\n\n# For Windows
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Setup Instructions (in the human-use repository)

1. Create and activate a virtual environment:

uv venv\n\n# On Unix/macOS
source .venv/bin/activate\n\n# On Windows
.venv\Scripts\activate

2. Install dependencies:

uv sync

Run the application

streamlit run app.py

Troubleshooting

If you encounter issues, with the dependencies make sure that "which python" and "which streamlit" are the same path. If they are not the same path, run "python -m streamlit run app.py" instead of "streamlit run app.py". If UV is not found, make sure you close all terminals and editors, then re-open a new one and try again.