README.md

metadata

title: Mcp Sentiment Analysis
emoji: 📊
colorFrom: yellow
colorTo: blue
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
pinned: false

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

MCP Sentiment (Gradio MCP Server)

Run a local MCP server that exposes a simple sentiment analysis tool via Gradio.

Prerequisites

• Python 3.10–3.12 recommended. Python 3.13 may work but some packages may lag support.
• pip (or conda) available in your shell
• Optional (for clients without SSE): Node.js with npx for mcp-remote

Install

# From repo root or this folder
pip install -r mcp-sentiment/requirements.txt

If you hit Python 3.13 issues, create a 3.11 env:

conda create -n mcp-sentiment python=3.11 -y
conda activate mcp-sentiment
pip install -r mcp-sentiment/requirements.txt

Run the server

cd mcp-sentiment
python app.py

You should see URLs like:

• Streamable HTTP URL: http://127.0.0.1:7860/gradio_api/mcp/
• [Deprecated] SSE URL: http://127.0.0.1:7860/gradio_api/mcp/sse

Connect a client

If your client supports HTTP/streaming MCP

Point it directly at the HTTP URL printed by Gradio (example above).

If your client only supports stdio (no SSE/HTTP)

Use mcp-remote to bridge HTTP/SSE to stdio.

1. Install Node (if you don’t have it):

# Install nvm
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.1/install.sh | bash
export NVM_DIR="$HOME/.nvm"; . "$NVM_DIR/nvm.sh"
# Install Node LTS (includes npx)
nvm install --lts
nvm use --lts

2. Configure your MCP client. Example config block:

{
  "mcpServers": {
    "gradio": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://localhost:7860/gradio_api/mcp/sse"
      ]
    }
  }
}

Tip: If your Gradio prints a non-SSE “Streamable HTTP URL”, you can use that instead:

"args": ["mcp-remote", "http://localhost:7860/gradio_api/mcp/"]

3. Restart your MCP client and look for a server named "gradio" with a sentiment tool.

Smoke test (optional)

Show available tools using the Inspector:

npx -y @modelcontextprotocol/inspector --command "npx" -- "mcp-remote" "http://localhost:7860/gradio_api/mcp/"

Notes

• The function sentiment_analysis returns a JSON string. Gradio uses the docstring to derive the tool schema exposed over MCP.
• If you change ports, update the URL in your MCP client config accordingly.