AutoMem

Bring persistent memory to your AI by connecting MCP clients to a graph-vector memory service. AutoMem MCP provides memory that survives conversations, sessions, and devices, enabling your AI to recall decisions, styles, and context across tools and platforms.

How to use

You connect an MCP client to the AutoMem memory service to enable persistent memory across your AI tools. After setup, your AI can recall past decisions, coding styles, and contextual preferences regardless of which platform you are using. Use the HTTP remote memory service for remote MCP or run a local AutoMem service for development. Once connected, memory operations such as storing, recalling, and associating memories become available to every MCP-enabled tool you choose to integrate.

How to install

Prerequisites you need before installation: Node.js and npm (for running MCP client setup), Git (for cloning the memory service during local development), and a network connection to install dependencies and run services.

Step 1. Set up the AutoMem memory service (local development) open a terminal and run the following commands to start a local memory backend.

# Clone the AutoMem service repository
git clone https://github.com/verygoodplugins/automem.git
cd automem

# Start the local development service (serves at http://localhost:8001)
make dev

Additional setup and usage notes

Step 2. Install the MCP client for AutoMem. This enables your MCP-enabled AI platforms to talk to the memory service and store or recall memories.

Run the guided setup to configure your first MCP client. This creates environment configuration and prints platform-specific snippets you can paste into your client.

# Guided setup for MCP client
npx @verygoodplugins/mcp-automem setup

During setup you will specify the AutoMem Endpoint. For a local deployment, use http://localhost:8001. If you deploy AutoMem to a remote host, use the provided URL for that host. You may optionally supply an API key if your deployment requires authentication.

Examples of how memory helps you code

• Your assistant remembers your preferred coding style, variable names, and comment conventions across projects.

• When deciding between technologies, memory-based reasoning surfaces past preferences and project-specific patterns to guide choices.

Available tools

store_memory

Save memories with content, tags, importance, and metadata for long-term retrieval.

recall_memory

Search memories with query support, including multi-query, tags, time filters, and graph-based expansion.

associate_memories

Create named relationships between memories to model complex connections.

update_memory

Modify existing memories to reflect new information or corrections.

delete_memory

Remove memories from the store when they are no longer relevant.

check_database_health

Monitor the health and status of the memory service.

multi_hop Recall

Perform multi-hop reasoning to answer complex questions by following related memories.

context_aware coding

Recall prioritizes language style and coding patterns to match your preferences.