Puchai

MCP sever for PuchAI hackathon

python

0

GitHub Stars

python

Language

7 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "nikh7l-puchai": {
      "command": "python",
      "args": [
        "mcp_starter.py"
      ],
      "env": {
        "MY_NUMBER": "YOUR_PHONE_NUMBER",
        "AUTH_TOKEN": "YOUR_AUTH_TOKEN"
      }
    }
  }
}

MCP enables your AI assistant to safely connect to external tools and data sources, giving it powerful capabilities while maintaining security. This starter provides a bearer-token MCP server for Puch AI with built-in text and image tools and a per-user task example to help you get started quickly.

How to use

You will run a local MCP server that Puch AI can connect to over HTTPS. Use the Bearer token variant to authenticate requests. The server exposes text and image processing tools and demonstrates per-user data scoping.

How to install

Prerequisites you need before starting the server:

• Python 3.11 or higher installed on your system.

Step 1 Install dependencies and set up the environment

# Create virtual environment
uv venv

# Install all required packages
uv sync

# Activate the environment
source .venv/bin/activate

Step 2 Set up environment variables

Create a .env file in the project root by copying the example and then editing it to include your credentials.

# Copy the example file
cp .env.example .env

# Edit .env to include your details
AUTH_TOKEN=your_secret_token_here
MY_NUMBER=919876543210

Step 3 Run the server

cd mcp-bearer-token
python mcp_starter.py

You should see a message like: 🚀 Starting MCP server on http://0.0.0.0:8086

## Step 4 Make it public (required by Puch)

To allow Puch AI to reach your server over HTTPS, expose your local instance with a tunnel or deploy it to the cloud.

## Option A: Using ngrok (Recommended)

Follow these steps to expose your local server securely.

Install ngrok (download from ngrok.com)

Get authtoken from the Ngrok dashboard and configure

ngrok config add-authtoken YOUR_AUTHTOKEN

Start the tunnel on the MCP port

ngrok http 8086

## Option B: Deploy to Cloud

You can also deploy to cloud services that host actions for MCP servers.

## Available tools

### text\_tool

Text input/output tool performing echo-style processing on incoming text data.

### image\_tool

Image input/output tool that can apply transformations (for example, convert to black & white) on images.

### task\_manager

Demo MCP server that illustrates using a per-user identifier (puch\_user\_id) to scope tasks and data per user.