hmem — Humanlike Memory for AI Agents

local MCP server

Installation

Terminal

claude mcp add hmem-mcp -e HMEM_STORAGE_DIR=... -- npx -y hmem-mcp

Run in your project folder. Add --scope user to use it everywhere.

Claude Desktop, Cursor & other clients

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

{
  "mcpServers": {
    "hmem-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "hmem-mcp"
      ],
      "env": {
        "HMEM_STORAGE_DIR": "..."
      }
    }
  }
}

Fill in the environment variables with your own credentials.

Summary

Persistent 5-level hierarchical memory for AI agents. SQLite-backed, lazy-loaded.

Questions

hmem — Humanlike Memory for AI Agents is a local MCP server distributed as the npm package hmem-mcp.

Run: claude mcp add hmem-mcp -e HMEM_STORAGE_DIR=... -- npx -y hmem-mcp

Run: codex mcp add hmem-mcp --env HMEM_STORAGE_DIR=... -- npx -y hmem-mcp

Similar MCP servers

NameStars

Waggle-mcpio.github.Abhigyan-Shekhar/Waggle-mcp44
enquire-mcpio.github.oomkapwn/enquire-mcp33
genesys-memoryio.github.Astrix-Labs/genesys-memory29
SAME - Stateless Agent Memory Engineio.github.sgx-labs/same23
LoreConvoio.github.labyrinth-analytics/loreconvo10
ori-memoryio.github.aayoawoyemi/ori-memory325
PMB AIio.github.oleksiijko/pmb-ai280
YourMemoryio.github.sachitrafa/yourmemory270