Integrations

Uses FastAPI to create a REST API server that handles requests for managing conversation context and interacting with language models.
Integrates with Google Gemini API to enable context-aware conversations with the language model, allowing the system to maintain conversation history across multiple requests.

MCP: Model Context Protocol

Project Description

MCP (Model Context Protocol) is a system for managing context when interacting with language models (LLM). The system allows you to save the context of dialogs between requests, which allows the language model to "remember" previous interactions.

Peculiarities

Maintaining and managing context between requests
Integration with Google Gemini API
Possibility of using multiple independent sessions
Switching between sessions using identifiers
Local storage of dialog history

Technologies

Python 3
FastAPI
Gemini API
Uvicorn

Installation

Clone repository:

git clone https://github.com/Alex-Replicator/mcp-letta.git
cd mcp-letta

Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate  # для Linux/Mac
# или
venv\Scripts\activate  # для Windows

Install dependencies:

pip install -r requirements.txt

Create a .env file and add the API key:

GEMINI_API_KEY=ваш_ключ_gemini_api

Usage

Starting the server

python mcp_server.py

The server will be available at: http://localhost:9999

Using the client

Send request (will create a new session):

python mcp_client.py --new "Ваш запрос к модели"

Continue the dialogue in the same session:

python mcp_client.py "Ваш следующий запрос"

Use a specific session by ID:

python mcp_client.py --id "id_сессии" "Ваш запрос"

Show list of all sessions:

python mcp_client.py --list

Project structure

mcp_server.py - Main MCP server
mcp_client.py - Client for interaction with the server
requirements.txt - Project dependencies

License

MIT

Author

Alex Replicator - alex.replicator@gmail.com

GitHub: https://github.com/Alex-Replicator/

This server cannot be installed

-

security - not tested

F

license - not found

-

quality - not tested

A system that manages context for language model interactions, allowing the model to remember previous interactions across multiple independent sessions using Gemini API.