emojikey-server MCP Server
MCP Protocol for persisting LLM interaction style as emojikeys
This server allows LLMs to maintain consistent interaction styles across conversations using emoji-based context keys ("emojikeys").
Emojikeys are stored online, so you can use them accross devices and applications. No user information is stored other than the emojikeys.
If there is interest in a local-only version, or other suggestions, please reach out and let me know. Actively developing this concept now.
📝 NoteUsage note: The first time you use the tool in Claude desktop, tell Claude to "Set emojikey" then next time you start a conversation, he will automatically use this key. You can ask to set vibe, or show emojikey history as well. Have fun!
⚠️ WarningThis is a beta version, more features are planned, so the API may change.
Features
Emojikey Management
- Get current emojikey for a user/model combination
- Set new emojikeys during conversations
- Retrieve emojikey history
- Automatic API key generation and validation
Tools
initialize_conversation
- Get current emojikey at start of conversationget_emojikey
- Retrieve current emojikeyset_emojikey
- Update the emojikeyget_emojikey_history
- View previous emojikeys
Installation
Get your API key from emojikey.io
Add the server config to Claude Desktop:
On MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json
On Windows: %APPDATA%/Claude/claude_desktop_config.json
{
"mcpServers": {
"mcp-server-emojikey": {
"command": "npx",
"args": ["@identimoji/mcp-server-emojikey"],
"env": {
"EMOJIKEYIO_API_KEY": "your-api-key-goes-here", //get it from emojikey.io
"MODEL_ID": "Claude-3-5-Sonnet-20241022"
}
}
}
}
Development
Install dependencies:
npm install
Build the server:
npm run build
For development with auto-rebuild:
npm run watch
Test the server:
npm run test
Environment Variables
EMOJIKEYIO_API_KEY
- Your emojikey.io API keyMODEL_ID
- Identifier for the LLM model (e.g., "Claude-3-5-Sonnet-20241022")
Debugging
Since MCP servers communicate over stdio, debugging can be challenging. We recommend:
- Using the test script:
npm run test
- Using the MCP Inspector:
npm run inspector