Memory-IA

Servidor MCP (Model Context Protocol) em Python para agente de IA com memória persistente, integrando FastAPI, LangGraph, SQLite e modelos locais via Ollama, pronto para uso em VS Code, Gemini-CLI e outros clientes MCP.

python

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": {
    "helciocosta-memory-ia-mcp": {
      "command": "python",
      "args": [
        "src/mcp_server.py"
      ],
      "env": {
        "DEBUG": "False",
        "AGENT_PORT": "8000",
        "OLLAMA_URL": "http://localhost:11434"
      }
    }
  }
}

Memory-IA MCP Server is a persistent, memory-enabled MCP server that powers AI agents in VS Code, Gemini-CLI, and terminal workflows. It maintains conversation context across interactions using SQLite, runs locally with an Ollama-backed local LLM, and exposes a standard JSON-RPC MCP interface for easy integration with your tools.

How to use

You can interact with Memory-IA MCP Server from various clients. Use a terminal-based workflow by sending JSON-RPC requests to the server’s standard input/output, or connect via VS Code or Gemini-CLI. The server keeps memory of past chats to provide contextual responses and supports local model execution through Ollama.

How to install

Prerequisites you need before installing are Python 3.12 and Git. You will also use a Python virtual environment to isolate dependencies.

# 1. Clone the repository
cd ~
git clone https://github.com/seu-usuario/memory-ia-mcp.git
cd memory-ia-mcp

# 2. Create a virtual environment
python3 -m venv memorivenv
source memorivenv/bin/activate

# 3. Install dependencies
pip install -r requirements.txt

# 4. Run the MCP server
./run_mcp.sh

Configuration and startup notes

Configure environment variables in a local .env file to control connections and server behavior. The following variables are commonly used by Memory-IA MCP Server.

OLLAMA_URL=http://localhost:11434
AGENT_PORT=8000
DEBUG=False

Service and maintenance

Manage the server as a systemd service to enable automatic restarts and centralized logging.

# Status
sudo systemctl status memory-ia-mcp.service

# Logs
sudo journalctl -u memory-ia-mcp -f

# Control
sudo systemctl restart memory-ia-mcp
sudo systemctl stop memory-ia-mcp
sudo systemctl start memory-ia-mcp

Notes on structure

The project organizes source, configuration, and tooling as follows. This layout helps you understand where to find the MCP server script and configuration.

Available tools

memory_chat

Chat with persistent memory enabled, retains context across interactions.

run_ollama

Run an Ollama model locally to provide local LLM inference.

agent_health

Check the health/status of the agent and its memory store.

Built by

VeilStrat

AI signals for GTM teams