Agentbay

Name: Agentbay
Rating: 3.5 (1 reviews)
Author: thomasjumper

By thomasjumper 👁 7 views ▲ 0 votes

OpenClaw plugin for AgentBay — persistent memory with auto-recall and auto-capture

Homepage GitHub

Install

pip install agentbay

Configuration Example

{
  "openclaw-mem": {
    "enabled": true,
    "config": {
      "apiKey": "ab_live_your_key"
    }
  }
}

README

# @agentbay/openclaw-mem

Persistent memory for OpenClaw agents. Auto-recall and auto-capture — zero manual work.

## Install

```bash
openclaw plugins install @agentbay/openclaw-mem
```

Or manually add to your `openclaw.json`:

```json
{
  "openclaw-mem": {
    "enabled": true,
    "config": {
      "apiKey": "ab_live_your_key"
    }
  }
}
```

## What It Does

- **Before each response**: searches AgentBay for relevant memories and injects them
- **After each response**: extracts learnings and stores them for next time
- Both run silently. No manual store/recall needed.

## Get an API Key

1. pip install agentbay && python -c "from agentbay import AgentBay; AgentBay().login()"
2. Or sign up at https://www.aiagentsbay.com/register

## Config Options

| Option | Default | Description |
|--------|---------|-------------|
| apiKey | required | Your AgentBay API key |
| projectId | auto | Project for memory storage |
| autoRecall | true | Inject memories before responses |
| autoCapture | true | Extract learnings after responses |
| topK | 5 | Max memories per recall |
| userId | "default" | User ID for scoped memory |

voice