OpenClaw MongoDB Semantic Memory

OpenClaw Memory — Agent Skill

MongoDB-backed long-term memory with Voyage AI semantic search

When to Use

Use OpenClaw Memory when:

NOT for:

Available Tools

memory_search

Semantically search long-term memory. Use this to recall prior decisions, preferences, context, or facts.

memory_search({
  query: "What did we decide about the database schema?",
  maxResults: 6  // optional, default: 6
})

Returns: Array of memories with similarity scores, text, tags, and metadata.

When to use:

Example output:

{
  "results": [
    {
      "id": "507f1f77bcf86cd799439011",
      "text": "Decided to use MongoDB for vector storage with Atlas Search",
      "score": 0.89,
      "tags": ["decision", "database"],
      "createdAt": "2026-02-20T14:30:00Z"
    }
  ]
}

memory_remember

Store a fact, decision, preference, or important context in long-term memory.

memory_remember({
  text: "User prefers TypeScript over JavaScript for new projects",
  tags: ["preference", "programming"],  // optional
  ttl: 2592000  // optional, 30 days default
})

Returns: Stored memory ID and confirmation.

When to use:

Best practices:

memory_get

Read a specific memory file from the workspace. Use memory_search for semantic recall; use this for targeted file reads.

memory_get({
  path: "MEMORY.md",
  from: 1,      // optional, starting line
  lines: 50     // optional, number of lines
})

Returns: File contents (text).

When to use:

memory_forget

Delete a specific memory by ID. Use memory_search first to find the memory ID.

memory_forget({
  memoryId: "507f1f77bcf86cd799439011"
})

Returns: Confirmation or error.

When to use:

memory_list

Browse stored memories by recency or tag.

memory_list({
  tags: "decision,database",  // optional, comma-separated
  limit: 10,                  // optional, default: 10
  sort: "desc"                // optional, "desc" or "asc"
})

Returns: Array of memories with metadata (no similarity scores).

When to use:

memory_status

Check memory system health and stats.

memory_status()

Returns: Daemon status, MongoDB connection, Voyage AI status, total memories, uptime.

When to use:

Configuration

Memory tools connect to a daemon at http://localhost:7654 by default. Configuration is set in ~/.openclaw/openclaw.json:

{
  plugins: {
    entries: {
      "openclaw-memory": {
        enabled: true,
        config: {
          daemonUrl: "http://localhost:7654",
          agentId: "openclaw",
          maxResults: 6,
          minScore: 0.5,
          defaultTtl: 2592000  // 30 days
        }
      }
    }
  }
}

Automatic Memory Capture

OpenClaw Memory includes lifecycle hooks that capture memories automatically:

auto-remember Hook

Limits: Max 5 extractions per message, min 10 chars, deduplicates.

session-to-memory Hook

memory-bootstrap Hook

memory-enriched-tools Hook

To disable hooks: Set hooksEnabled: false in plugin config.

Workflow Examples

Example 1: Recall Prior Decision

User asks: "What did we decide about the API authentication?"

Agent response: 1. Call memory_search({ query: "API authentication decision" }) 2. Review results 3. Answer based on stored memory 4. If no results, say "I don't have any memory of that decision"

Example 2: Store Preference

User says: "I prefer Material UI over Tailwind for all React projects"

Agent response: 1. Acknowledge the preference 2. Call memory_remember({ text: "User prefers Material UI over Tailwind for React projects", tags: ["preference", "ui"] }) 3. Confirm it's stored: "Got it, I'll remember that preference"

Example 3: Check Before Recommending

User asks: "What CSS framework should we use?"

Agent response: 1. Call memory_search({ query: "CSS framework preference" }) 2. If match found: "You previously preferred Material UI over Tailwind" 3. If no match: Provide recommendation based on context

Example 4: Session Continuity

New session starts:

1. memory-bootstrap hook auto-runs 2. Loads recent preferences, decisions, project context 3. Agent has continuity without user repeating everything

Tips & Best Practices

Do:

Don't:

Search Tips:

TTL Guidelines:

Troubleshooting

"Memory daemon not reachable"

"No memories found"

"Memory search returns irrelevant results"

"Tools not available"

Advanced Features

Web Dashboard

Full installation includes a web dashboard at http://localhost:3002:

Reflection Pipeline

9-stage processing pipeline for:

Trigger reflection:

curl -X POST http://localhost:7654/reflect \
  -H "Content-Type: application/json" \
  -d '{"agentId":"openclaw"}'

Graph Relationships

Memories can be connected via edges:

Access via web dashboard at /graph.

Requirements

Installation

# Install plugin
openclaw plugins install openclaw-memory
Start daemon
cd openclaw-memory
pnpm install && pnpm dev:daemon
Or use Docker
docker compose up -d

Summary

OpenClaw Memory gives agents persistent, searchable memory across sessions:

1. Search semantically with memory_search 2. Store facts with memory_remember 3. Automatic capture via lifecycle hooks 4. MongoDB-backed with Voyage AI embeddings 5. Web dashboard for visualization and management

Use it to build agents that remember, learn, and improve over time. 🧠

Version: 0.2.1 Author: Michael Lynn License: MIT Repository: https://github.com/mrlynn/openclaw-mongodb-memory

Memory tools connect to a daemon at http://localhost:7654 by default. Configuration is set in ~/.openclaw/openclaw.json:

{
  plugins: {
    entries: {
      "openclaw-memory": {
        enabled: true,
        config: {
          daemonUrl: "http://localhost:7654",
          agentId: "openclaw",
          maxResults: 6,
          minScore: 0.5,
          defaultTtl: 2592000  // 30 days
        }
      }
    }
  }
}

"Memory daemon not reachable"

"No memories found"

"Memory search returns irrelevant results"

"Tools not available"