Tools
TencentAgentMemoryBridge
Memory bridges for AI Agents — wraps TencentDB Agent Memory (4-tier local long-term memory) into MCP Server & OpenClaw plugin, enabling CodeBuddy, Claude Code, and OpenClaw agents with persistent memory capabilities.
Configuration Example
// codebuddy.json
{
"mcpServers": {
"agent-memory": {
"command": "npx",
"args": ["@tencent-agent-memory/mcp-bridge"],
"env": {
"MEMORY_DATA_DIR": "./.agent-memory"
}
}
}
}
README
# TencentAgentMemoryBridge
围绕 [TencentDB Agent Memory](https://github.com/TencentCloud/TencentDB-Agent-Memory) 构建的记忆桥梁生态——将 4 层长期记忆能力接入不同的 Agent 平台。
## 项目目标
**不造轮子**。我们复用腾讯开源的 TencentDB Agent Memory(L0 对话 → L1 原子事实 → L2 场景 → L3 画像)作为记忆引擎,通过以下桥梁让不同生态的 Agent 获得长期记忆能力:
- **MCP Bridge** — 通过 MCP(Model Context Protocol)协议,让 CodeBuddy、Claude Code 等工具获得长期记忆
- **OpenClaw Plugin** — 增强 OpenClaw 生态的多 Agent 共享记忆能力
## 快速开始
### MCP Bridge(适用于 CodeBuddy / Claude Code)
```bash
npx @tencent-agent-memory/mcp-bridge
```
然后在工具中配置 MCP Server:
<details>
<summary>CodeBuddy 配置</summary>
```jsonc
// codebuddy.json
{
"mcpServers": {
"agent-memory": {
"command": "npx",
"args": ["@tencent-agent-memory/mcp-bridge"],
"env": {
"MEMORY_DATA_DIR": "./.agent-memory"
}
}
}
}
```
</details>
<details>
<summary>Claude Code 配置</summary>
```jsonc
// ~/.claude/settings.json
{
"mcpServers": {
"agent-memory": {
"command": "npx",
"args": ["@tencent-agent-memory/mcp-bridge"],
"env": {
"MEMORY_DATA_DIR": "./.agent-memory"
}
}
}
}
```
</details>
## 项目结构
```
tencent-agent-memory-bridge/
├── packages/
│ ├── mcp-bridge/ # MCP Server(CodeBuddy + Claude Code)
│ └── openclaw-plugin/ # OpenClaw 插件改造
├── examples/ # 配置示例
├── docs/ # 设计文档
└── package.json
```
## 上游依赖
- [TencentDB Agent Memory](https://github.com/TencentCloud/TencentDB-Agent-Memory) — 腾讯开源的 4 层本地长期记忆系统(MIT 协议)
## License
MIT
tools
Comments
Sign in to leave a comment