Connects to Kogna's multi-agent AI avatar system hosted on Railway, providing tools for managing conversations, switching between avatars and rooms, and accessing conversation history through the Kogna backend.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Kogna MCP Serverlist available avatars and their specialties"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
title: Kogna MCP Server emoji: 🤖 colorFrom: blue colorTo: purple sdk: docker pinned: false license: mit
Kogna MCP Server
A Model Context Protocol (MCP) server that provides tools for interacting with Kogna's multi-agent AI avatar system.
Installation
Install pipx (if you don't have it)
# macOS
brew install pipx
# Other systems
pip install pipx
# Add pipx to your PATH (if not already done)
echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.zshrc
source ~/.zshrcInstall kognamcp
pipx install kognamcpTesting the Installation
1. Test the command is available
kognamcp --helpYou should see:
usage: kognamcp [-h]
Kogna MCP Bridge - Forwards MCP requests to the Kogna HTTP server
options:
-h, --help show this help message and exit
Examples:
# Run as MCP server (reads from stdin, writes to stdout)
kognamcp
# Test with echo (for development)
echo '{"jsonrpc": "2.0", "method": "initialize", "id": 1}' | kognamcp2. Test the MCP bridge connection
echo '{"jsonrpc": "2.0", "method": "initialize", "id": 1}' | kognamcpYou should get a response like:
{"jsonrpc": "2.0", "id": 1, "result": {"protocolVersion": "2024-11-05", "capabilities": {"tools": {}}, "serverInfo": {"name": "kognamcp-server", "version": "1.0.0"}}}Configuration
Find the kognamcp path
First, find the full path to the kognamcp command:
which kognamcpThis will output something like /Users/username/.local/bin/kognamcp
Configure your MCP client
Add this configuration to your MCP client:
{
"mcpServers": {
"kognamcp": {
"command": "FULL_PATH_FROM_WHICH_COMMAND"
}
}
}Replace
For example, if which kognamcp returns /Users/john/.local/bin/kognamcp, your config should be:
{
"mcpServers": {
"kognamcp": {
"command": "/Users/john/.local/bin/kognamcp"
}
}
}Troubleshooting
"Command not found: kognamcp"
Make sure the installation completed successfully:
pipx list | grep kognamcpAdd pipx to your PATH:
echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.zshrc source ~/.zshrcRestart your terminal/IDE
"spawn kognamcp ENOENT" or similar errors
This happens when your MCP client can't find the kognamcp command. The solution is to use the full path:
Find the full path:
which kognamcpCopy the exact output and use it in your MCP configuration
Restart your MCP client
Connection errors
The bridge connects to https://kogna.up.railway.app/mcp. Make sure you have internet access.
Available Tools
start_kogna_conversation- Start a new conversation with Kogna avatarssend_kogna_message- Send a message to the current conversationlist_kogna_avatars- List available avatars and their specialtieslist_kogna_rooms- List available rooms and their descriptionsswitch_kogna_avatar- Switch to a different avatarswitch_kogna_room- Switch to a different roomget_kogna_conversation_history- Get conversation historyget_kogna_system_info- Get system information
Usage
Once configured, you can use Kogna through your MCP client:
"Start a Kogna conversation"
"List the available Kogna avatars"
"Switch to the business room"
"Ask the strategist avatar about market positioning"
Requirements
Python 3.8+
Internet connection (connects to Kogna backend)
License
MIT License