IAI MCP

v1.0.0Knowledge & Memorystable

A local MCP server that gives AI assistants a long-term memory by capturing sessions verbatim and surfacing relevant context automatically.

ai-agentsai-memoryclaudeclaude-codeembeddings
Share:
124
Stars
0
Downloads
0
Weekly
0/5

What is IAI MCP?

IAI MCP is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to local mcp server that gives ai assistants a long-term memory by capturing sessions verbatim and surfacing relevant context automatically.

A local MCP server that gives AI assistants a long-term memory by capturing sessions verbatim and surfacing relevant context automatically.

This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • A local MCP server that gives AI assistants a long-term memo

Use Cases

Capture AI session history for long-term episodic memory.
Surface relevant context automatically using embeddings.
CodeAbra

Maintainer

LicenseMIT License
Languagepython
Versionv1.0.0
UpdatedMay 20, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx iai

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use IAI MCP

IAI MCP (iai-personal-memory-engine) is a local MCP server that gives AI assistants persistent, long-term episodic memory by automatically capturing every session turn verbatim and surfacing relevant past context at the start of each new session. It uses a native Rust embedding engine with the bge-small-en-v1.5 model and LanceDB for local vector storage, so all data stays on your machine. Developers and power users who work daily with Claude Code or Codex CLI benefit most — the server eliminates the need to re-explain project context every session and builds a searchable history of past decisions and conversations.

Prerequisites

  • macOS (Apple Silicon tested; Linux may work); Python 3.11 or 3.12
  • Node.js 18 or later
  • Rust toolchain (installed via rustup) for building the native engine
  • An MCP-compatible client such as Claude Code or Codex CLI
  • ~500 MB free disk space for the LanceDB store and model weights
1

Clone and install the Python package

Clone the repository, create a Python virtual environment with Python 3.12, and install the package. The installer automatically compiles the native Rust engine via setuptools-rust.

git clone https://github.com/CodeAbra/iai-personal-memory-engine.git
cd iai-personal-memory-engine
python3.12 -m venv .venv && source .venv/bin/activate
pip install .
2

Build the MCP Node.js wrapper

The MCP transport layer is a small Node.js wrapper around the Python engine. Build it and then install the background daemon that keeps the memory engine running between sessions.

cd mcp-wrapper && npm install && npm run build && cd ..
iai-mcp daemon install
3

Install automatic capture hooks

Hook scripts patch your Claude Code (or Codex CLI) settings so every conversation turn is captured automatically without any manual commands. Run the appropriate variant for your client.

# For Claude Code only
iai-mcp capture-hooks install

# For Codex CLI only
iai-mcp capture-hooks install --target codex

# For both clients
iai-mcp capture-hooks install --target all
4

Register the MCP server with Claude Code

Add the MCP server to Claude Code using an absolute path to the built Node.js wrapper. Tilde expansion does not work in config files — use the full path.

claude mcp add iai-mcp -- node "/absolute/path/to/iai-personal-memory-engine/mcp-wrapper/dist/index.js"
5

Verify the installation

Run the built-in health checker which performs 25 diagnostic checks, then confirm the daemon is active and that the first capture log exists after starting a Claude Code session.

iai-mcp doctor
iai-mcp daemon status

IAI MCP Examples

Client configuration

Add this block to ~/.claude.json (Claude Code) to register the MCP server. Set IAI_MCP_PYTHON to your venv Python and IAI_MCP_STORE to the desired data directory.

{
  "mcpServers": {
    "iai-mcp": {
      "command": "node",
      "args": ["/absolute/path/to/iai-personal-memory-engine/mcp-wrapper/dist/index.js"],
      "env": {
        "IAI_MCP_PYTHON": "/absolute/path/to/iai-personal-memory-engine/.venv/bin/python",
        "IAI_MCP_STORE": "/Users/yourname/.iai-mcp"
      }
    }
  }
}

Prompts to try

After installation, memory capture and recall are fully automatic. These prompts interact with the memory system or inspect its state.

- "What have we discussed about the authentication module in past sessions?"
- "Summarize the architectural decisions we made last week."
- "What was the last bug we fixed together?"
- "Show me the topology of my memory store."

Troubleshooting IAI MCP

iai-mcp doctor reports Rust engine build failure

Ensure the Rust toolchain is installed (run 'rustup show') and that you are using Python 3.11 or 3.12. Then re-run 'pip install .' inside the activated venv.

Memory is not recalled at session start

Confirm the daemon is running with 'iai-mcp daemon status'. If it is stopped, run 'iai-mcp daemon install' again and check logs at ~/.iai-mcp/logs/ for errors.

Config paths with ~ or $HOME are not expanded

Use fully qualified absolute paths in claude_desktop_config.json or ~/.claude.json. The MCP host does not perform shell expansion on config values.

Frequently Asked Questions about IAI MCP

What is IAI MCP?

IAI MCP is a Model Context Protocol (MCP) server that local mcp server that gives ai assistants a long-term memory by capturing sessions verbatim and surfacing relevant context automatically. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install IAI MCP?

Follow the installation instructions on the IAI MCP GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with IAI MCP?

IAI MCP works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is IAI MCP free to use?

Yes, IAI MCP is open source and available under the MIT License license. You can use it freely in both personal and commercial projects.

Browse More Knowledge & Memory MCP Servers

Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "iai": { "command": "npx", "args": ["-y", "iai"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use IAI MCP?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides