🔍
—
🔍

Related Memories

Shared Memory

🔍

Skill Files
SKILL.md Content
Version History
Related Tasks
Range
-
Total Memories
-
Writes Today
-
Sessions
-
Embeddings

📊 Memory Writes per Day

⚡ Tool Response Time (per minute avg)

| –
🧠
AI Models
Configure embedding, summarizer and skill evolution models
🔔 模型配置重要提醒
  • 嵌入模型 (Embedding)：插件内置模型规模较小。为获得更精准的记忆检索体验，强烈建议配置 bge-m3 等专业嵌入模型。
  • 摘要模型 (Summarizer)：此项为必填项，否则无法自动提取记忆摘要。建议配置非思考型大模型，以保障处理速度和流畅度。
  • 技能模型 (Skill Evolution)：用于自动提取可复用技能。建议配置思考型大模型，以获得最佳的生成效果和稳定性。
📡 Embedding Model
Vector embedding model for memory search and retrieval
📝 Summarizer Model
LLM for memory summarization, deduplication and analysis
🔧 Skill Evolution
Auto-extract reusable skills from conversation patterns
Skill Dedicated Model
If not configured, the main Summarizer Model above will be used for skill generation. Configure a dedicated thinking model here for higher quality skill output.
✓ Saved
Some changes require restarting the OpenClaw gateway to take effect.
🌐
Team Sharing
Share memories, tasks and skills with your team
🚀 Get Started with Team Collaboration
MemOS supports team memory sharing. Choose one of the following options to enable collaboration, or continue using local-only mode.
🌐
Join a Remote Team
Your team already has a server running? Join it to share memories, tasks and skills with team members.
  1. Ask your team admin for the Server Address and Team Token
  2. Enable sharing above, select "Client" mode
  3. Fill in Server Address and Team Token, click "Test Connection"
  4. Click "Save & Apply" — the service restarts automatically (page refreshes)
🖥️
Start Your Own Team Server
Be the team server. Run it on this device so others can connect and share memories with you.
  1. Enable sharing above, select "Server" mode
  2. Set a team name, click "Save & Apply" — the service restarts automatically
  3. Share the Server Address and Team Token with your team members
  4. Approve join requests in the Admin Panel
Enable to share memories, tasks and skills with your team. When disabled, all features work normally in local-only mode.
✓ Saved
Some changes require restarting the OpenClaw gateway to take effect.
⚙️
General
System status, ports and telemetry
📊 Model Health
Loading model status...
Requires restart to take effect
Active tasks with no new messages beyond this duration will be automatically summarized and completed when the Tasks page is opened. Set to 0 to disable. Default: 4 hours.
Anonymous usage analytics to help improve the plugin. Only sends tool names, latencies, and version info. No memory content, queries, or personal data is ever sent.
✓ Saved

⚡ Team Admin Panel

Manage team members and shared resources

📥 Import OpenClaw Memory

Migrate your existing OpenClaw built-in memories and conversation history into this plugin. The import process uses smart deduplication to avoid duplicates.

Three ways to use:
① Import memories only (fast) — Click "Start Import" to quickly migrate all memory chunks and conversations. No task/skill generation. Suitable when you just need the raw data.
② Import + generate tasks & skills (slow, serial) — After importing memories, enable "Generate Tasks" and/or "Trigger Skill Evolution" below to analyze conversations one by one. This takes longer as each session is processed by LLM sequentially.
③ Import first, generate later (flexible) — Import memories now, then come back anytime to start task/skill generation. You can pause the generation at any point and resume later — it will pick up where you left off, only processing sessions that haven't been handled yet.