Getting started

Quickstart

Get MemoryLayer running in under 5 minutes. You need Docker installed.

Prerequisites

  • • Docker Desktop (Mac/Windows) or Docker Engine (Linux)
  • •A MemoryLayer API key — get one free
  • • Claude Desktop, Cursor, or any MCP-compatible agent

Installation

1
Pull and run the container
Terminal
docker run -d \
  --name memorylayer \
  --restart unless-stopped \
  -e MEMORY_API_KEY=sk-ml-your-key-here \
  -v memorylayer-data:/data \
  -p 8765:8765 \
  memorylayer/memorylayer:latest
The container starts the REST API on port 8765 and preloads the embedding model on first run.
2
Verify it's running
Terminal
curl http://localhost:8765/health
# → {"status":"ok","version":"1.0.0"}
3
Add to your MCP config
Open your Claude Desktop config file:
  • • macOS: ~/.claude/claude_desktop_config.json
  • • Windows: %APPDATA%\Claude\claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "memory": {
      "command": "docker",
      "args": [
        "exec", "-i", "memorylayer",
        "python", "-m", "src.mcp_server"
      ],
      "env": {
        "MEMORY_API_KEY": "sk-ml-your-key-here"
      }
    }
  }
}
Restart Claude Desktop after saving.
4
Test in Claude
Ask Claude:
"Store this in memory: my preferred coding style is
 TypeScript with strict mode and functional patterns."

"What do you remember about my coding preferences?"
Claude will use the memory_store and memory_search tools automatically.

Using with other agents

MemoryLayer also exposes a REST API for agents that don't support MCP:

Terminal
# Store a memory
curl -X POST http://localhost:8765/memories \
  -H "Content-Type: application/json" \
  -H "X-API-Key: sk-ml-your-key-here" \
  -d '{"content": "User prefers Python over Ruby", "namespace": "prefs"}'

# Search memories
curl "http://localhost:8765/memories/search?q=coding+preferences" \
  -H "X-API-Key: sk-ml-your-key-here"

Data persistence

All memories are stored in a Docker volume (memorylayer-data). To back up:

Terminal
docker run --rm \
  -v memorylayer-data:/data \
  -v $(pwd):/backup \
  alpine tar czf /backup/memorylayer-backup.tar.gz /data

Next steps

MCP Tools Reference — all available tools and parameters →GitHub — source code and issues →