Install and configure the memory-tencentdb plugin for OpenClaw to retain and retrieve conversation information. Verify the setup through logs and searches, adjust memory settings, and troubleshoot missing recall or vector search results.
用于在 OpenClaw 环境中安装、配置并验证 @tencentdb-agent-memory/memory-tencentdb 插件。当用户提到"安装记忆插件""配置 memory-tencentdb""开启长期记忆/召回"或出现相关报错时应触发。
Review the instructions the agent will receive before you install.
在不依赖外部托管记忆服务的前提下，为 OpenClaw 提供可持续的本地长期记忆能力（L0→L1→L2→L3），并完成从安装、配置到验收的一次性闭环。
memory-tencentdb先确认基础版本满足要求：
>= 2026.3.13>= 22.16.0执行：
openclaw --version
node -v
若版本不满足，先升级再继续。
执行安装命令：
openclaw plugins install @tencentdb-agent-memory/memory-tencentdb
如已安装则执行更新：
openclaw plugins update memory-tencentdb
编辑 ~/.openclaw/openclaw.json，确保存在：
{
"memory-tencentdb": {
"enabled": true
}
}
说明：该插件支持零配置启动；不补充其它字段也能运行基础能力。
根据用户需求补充如下分组：
capture: 对话捕获与保留策略extraction: L1 提取与去重pipeline: L1→L2→L3 调度recall: 召回数量、阈值、策略persona: 场景与画像触发参数embedding: 向量检索配置（远端 OpenAI 兼容）推荐模板：
{
"memory-tencentdb": {
"capture": {
"enabled": true,
"excludeAgents": [],
"l0l1RetentionDays": 90,
"cleanTime": "03:00"
},
"extraction": {
"enabled": true,
"enableDedup": true,
"maxMemoriesPerSession": 10,
"model": "provider/model"
},
"pipeline": {
"everyNConversations": 5,
"enableWarmup": true,
"l1IdleTimeoutSeconds": 600,
"l2DelayAfterL1Seconds": 10,
"l2MinIntervalSeconds": 900,
"l2MaxIntervalSeconds": 3600,
"sessionActiveWindowHours": 24
},
"recall": {
"enabled": true,
"maxResults": 5,
"scoreThreshold": 0.3,
"strategy": "hybrid"
},
"persona": {
"triggerEveryN": 50,
"maxScenes": 15,
"backupCount": 3,
"sceneBackupCount": 10,
"model": "provider/model"
},
"embedding": {
"enabled": true,
"provider": "openai",
"baseUrl": "https://api.openai.com/v1",
"apiKey": "${EMBEDDING_API_KEY}",
"model": "text-embedding-3-small",
"dimensions": 1536,
"conflictRecallTopK": 5
}
}
}
embedding.provider = "none" 时，向量能力会禁用，仅保留关键词路径。provider（如 openai / deepseek），必须同时提供：
apiKeybaseUrlmodeldimensionsl0l1RetentionDays：
0 表示不清理0 时建议 >=31~2，需显式开启 allowAggressiveCleanup执行：
openclaw gateway restart
检查项：
[memory-tdai] 前缀~/.openclaw/state/memory-tdai/conversations/、records/、scene_blocks/、vectors.db执行一次最小对话回路并验证：
tdai_memory_searchtdai_conversation_searchopenclaw.json 中 memory-tencentdb.enabled 是否为 true，并确认已重启 Gateway。recall.enabled、scoreThreshold 是否过高。embedding 四元组（apiKey/baseUrl/model/dimensions）是否齐全。l0l1RetentionDays 与 allowAggressiveCleanup。~/.openclaw/openclaw.json，并再次重启 Gateway。apiKey 视为敏感信息；不在聊天、日志、截图中明文扩散。memory-tencentdb 对应配置段，避免覆盖用户其它插件配置。在结束任务前，必须同时满足：
openclaw.json 已存在有效 memory-tencentdb 配置[memory-tdai] 日志可见可在完成后向用户输出：
memory-tencentdb 安装与配置，并重启 Gateway。recall.scoreThreshold、pipeline.everyNConversations、persona.triggerEveryN 与 embedding 模型参数。No reviews yet. Tell others what you thought.
No login needed. Share how it worked for you, ask a question, or note where it did not run.
No comments yet. Be the first to write one.
Connect an AI agent client to Memory Proxy, which adds team memory, skills, and knowledge to model requests. The guided workflow checks existing settings, tests connectivity, and writes the appropriate client configuration.