OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Memory & RAG

Montycat MCP

Montycat MCP is a self-hosted Model Context Protocol server that connects AI clients to a shared, persistent memory store backed by a Montycat Semantic engine. Connected agents can write and retrieve memories across sessions using semantic, keyword, hybrid, exact-key and metadata search.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Montycat MCP — OpenViST project profile in Agent Memory & RAG.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Montycat MCP when an MCP-connected agent workflow needs shared memory that persists between sessions, especially when clients should retrieve facts by meaning as well as by exact keywords.

Who it is for

  • Developers configuring Claude, Codex, Cursor or another MCP client to use persistent memory.
  • Teams that want connected agents to share project decisions and context.
  • Operators who want the memory server and embeddings to run on their own machine.

Documented uses

  • Save decisions, preferences and project context beyond a single chat.
  • Retrieve a saved memory by meaning, exact terms or a combination of search methods.
  • Let multiple connected agents read and write to the same memory store.
  • Wait for another agent’s memory update without polling.
Before you choose

Requirements & limits

The repository identifies the project as MIT-licensed. Documented setup options include a Claude Desktop package, a Claude Code plugin installation, and a Python 3.10+ uvx route for Codex, Cursor and other MCP clients.

  • Search options include vector/semantic, BM25 keyword, hybrid, exact-key and metadata lookup.
  • The default setup starts a local semantic engine; an existing engine can be selected with MONTYCAT_URI.
  • A memory scope is a namespace, not a security boundary; the documentation recommends separate MCP servers with least-privilege credentials for isolation.
  • The documented Claude Desktop package requires no Python, while the uvx route requires Python 3.10 or later.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Connect a Codex client to Montycat MCP

Register the Montycat MCP server with Codex using the documented uvx command.

Before you start

  • Codex
  • Python 3.10 or later
  • A client configuration that supports MCP stdio servers

Steps

  1. Point the client’s stdio configuration at uvx montycat-mcp, or use the documented Codex command.

  2. Run codex mcp add montycat -- uvx montycat-mcp.

Expected outcome

Codex is configured with a Montycat MCP server entry.

Guide limitations

  • This setup recipe does not describe a memory write or retrieval workflow.
  • The default setup starts a local Montycat Semantic engine; connecting to an existing engine requires MONTYCAT_URI configuration.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
MontyGovernance/montycat-mcp · Official homepage title · retained in its source language
How we review sourcesSuggest a correction