An MCP server that connects AI agents with human intelligence through the Rapidata API, enabling artificial intelligence to ask real people questions and receive feedback on various tasks.

Installation

From Source Code

git clone https://github.com/RapidataAI/human-use.git
# Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh
# On Windows: powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
uv venv
source .venv/bin/activate
uv sync

Running the Application

streamlit run app.py

Configuration

Cursor

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

Available Tools

Tool	Description
get_free_text_responses	Requests real people to provide short text answers to a question
get_human_image_classification	Requests real people to classify images in a directory
get_human_image_ranking	Requests real people to rank images in a directory
get_human_text_comparison	Requests real people to compare two texts and choose which one is better

Features

• Connecting AI agents with human intelligence
• Getting feedback from people on text responses
• Image classification by humans
• Image ranking by humans
• Text comparison by humans
• Hosted version available at chat.rapidata.ai
• Interface based on a custom Streamlit application

Use Cases

Coming up with a cool car design

Finding the best slogan

Function naming

Ranking different image generation models

Notes

Paths must be ABSOLUTE paths. If you encounter dependency issues, make sure that 'which python' and 'which streamlit' point to the same path. The application includes a custom Streamlit interface created due to issues with other MCP clients, such as the Claude desktop app. For support, contact info@rapidata.ai.