MemoryStack Plugin for OpenClaw

Long-term memory for OpenClaw (formerly Clawdbot). Automatically remembers conversations, recalls relevant context, and builds a persistent memory layer — all powered by MemoryStack cloud.

Install

openclaw plugins install @memorystack/clawdbot-memorystack

Restart OpenClaw after installing.

Configuration

The only required value is your MemoryStack API key. Get one at memorystack.app/dashboard/api-keys.

Set it as an environment variable:

export MEMORYSTACK_API_KEY="ms_proj_..."

Or configure it directly in openclaw.json (or moltbot.json):

{
  "plugins": {
    "entries": {
      "clawdbot-memorystack": {
        "enabled": true,
        "config": {
          "apiKey": "${MEMORYSTACK_API_KEY}",
          "autoRecall": true,
          "autoCapture": true
        }
      }
    }
  }
}

Advanced options

| Key | Type | Default | Description | |-----|------|---------|-------------| | apiKey | string | - | Your MemoryStack API key (required) | | baseUrl | string | https://memorystack.app | API endpoint URL | | autoRecall | boolean | true | Inject relevant memories before every AI turn | | autoCapture | boolean | true | Automatically store conversation content after every turn | | maxRecallResults | number | 5 | Max memories injected into context per turn (1-20) | | debug | boolean | false | Verbose debug logs for API calls and responses |

Slash Commands

Use these commands directly in chat:

| Command | Description | |---|---| | /add <text> | Save something to long-term memory | | /search <query> | Search your memories | | /stats | View memory usage statistics |

CLI Commands

Manage memories from the terminal:

openclaw memorystack search "project alpha"
openclaw memorystack stats
openclaw memorystack add "The server IP is 10.0.0.5"
openclaw memorystack deleteall   # Destructive!

How it works

Once installed, the plugin works automatically with zero interaction:

• Auto-Recall — Before every AI turn, the plugin queries MemoryStack for relevant memories and injects them as context using semantic search.
• Auto-Capture — After every AI turn, the last user/assistant exchange is sent to MemoryStack for extraction and long-term storage.

Everything runs in the cloud. MemoryStack handles importance scoring, contradiction detection, consolidation, and reflection on its end.

Features

• Hybrid Search - Vector similarity + text search with RRF ranking
• Importance Scoring - Automatic ranking of what matters
• Contradiction Detection - Resolves conflicts and updates beliefs
• Memory Consolidation - Deduplication and merging (runs server-side)
• Temporal Decay - Simulates natural forgetting
• Reflection - Discovers patterns and generates insights

Links

• 🌐 Website: https://memorystack.app
• 📚 Documentation: https://memorystack.app/docs
• 🚀 Quick Start: https://memorystack.app/docs/quickstart
• 🔑 Get API Key: https://memorystack.app/dashboard/api-keys
• 💰 Pricing: https://memorystack.app/pricing

License

MIT