ClawMem
Shared memory for AI agents.
ClawMem gives OpenClaw, Claude Code, Codex, and Hermes a shared memory layer that persists across sessions, survives agent handoffs, keeps team context in sync, and stays inspectable, editable, and yours.
Choose your agent and copy the quick start.
Read https://clawmem.ai/SKILL.md and follow the instructions to install ClawMem, start recording chats, and build your memory graph in OpenClaw curl -fsSL https://raw.githubusercontent.com/clawmem-ai/clawmem-hermes-plugin/main/install.sh | bash Start a new Hermes session after the script finishes. If activation fails, run hermes memory setup. Verify anytime with hermes memory status.
claude plugin marketplace add https://github.com/clawmem-ai/clawmem-claude-code-plugin
claude plugin install clawmem-claude-code-plugin@clawmem claude codex plugin marketplace add clawmem-ai/clawmem-codex-plugin --ref main Codex now supports the official ClawMem plugin marketplace flow directly from GitHub.
codex plugin add clawmem@clawmem-ai Open a new Codex thread after install so the plugin skill and MCP tools load cleanly.
npx -y clawmem-mcp-server {
"mcpServers": {
"clawmem": {
"command": "npx",
"args": ["-y", "clawmem-mcp-server"]
}
}
} Team memory
One shared memory layer for the whole team
Updates captured by one teammate become usable by the next agent immediately.
From raw sessions to durable context
Capture the session
Every conversation is recorded automatically. Nothing falls through the cracks.
Store what matters
Key facts, preferences, and decisions are distilled into structured, durable memory records.
Recall it when needed
Before the next run, relevant memories are found and injected into context automatically.
Capture the session
Someone in your family tells their agent something important — it gets recorded in shared memory.
Store what matters
Key facts, preferences, and decisions are distilled into structured, durable memory records.
Recall it when needed
Your agent pulls in memories from family members, so you never miss what matters.
Capture the session
Every conversation is recorded automatically. Nothing falls through the cracks.
Store what matters
Key facts, preferences, and decisions are distilled into structured, durable memory records.
Recall it when needed
Before the next run, relevant memories are found and injected into context automatically.
Capture the session
A teammate tells their agent about an update — it's stored in shared team memory.
Store what matters
Key facts, preferences, and decisions are distilled into structured, durable memory records.
Recall it when needed
Your agent surfaces teammates' updates before you start — no Slack archaeology needed.
Memory should feel inspectable, not mystical
Every memory is a structured record you can see, search, edit, and trace back to its source. No hidden embeddings, no black boxes.
Visible
Browse all memories in a structured list. Nothing is hidden.
Searchable
Semantic search by meaning, not just keywords.
Editable
Fix, merge, or delete any memory before it shapes the next answer.
Traceable
Every memory links back to the exact session and turn that created it.
Under the hood
ClawMem sits between your agents and a structured, inspectable backend that keeps shared memory consistent across sessions.
GitHub-compatible API
Operate memory with tools you already know. REST, GraphQL, Git HTTP.
Searchable durable records
Vector-embedded memories stored in a persistent, queryable backend.
Automatic memory recall
Relevant context is retrieved and injected before every agent run.
With vs without ClawMem
Start with the agent you already use.
Copy the quick start for the agent you already use.
Read https://clawmem.ai/SKILL.md and follow the instructions to install ClawMem, start recording chats, and build your memory graph in OpenClaw curl -fsSL https://raw.githubusercontent.com/clawmem-ai/clawmem-hermes-plugin/main/install.sh | bash Start a new Hermes session after the script finishes. If activation fails, run hermes memory setup. Verify anytime with hermes memory status.
claude plugin marketplace add https://github.com/clawmem-ai/clawmem-claude-code-plugin
claude plugin install clawmem-claude-code-plugin@clawmem claude codex plugin marketplace add clawmem-ai/clawmem-codex-plugin --ref main Codex now supports the official ClawMem plugin marketplace flow directly from GitHub.
codex plugin add clawmem@clawmem-ai Open a new Codex thread after install so the plugin skill and MCP tools load cleanly.
npx -y clawmem-mcp-server {
"mcpServers": {
"clawmem": {
"command": "npx",
"args": ["-y", "clawmem-mcp-server"]
}
}
} Each tab gives you a copyable quick start for that agent.
Your memories are stored privately and belong entirely to you. Browse, export, or delete anytime.
Already installed? Go to Console to inspect your shared memory.