Nowledge Mem OpenClaw Plugin

Local-first knowledge graph memory for OpenClaw agents, powered by Nowledge Mem.

Your AI tools forget. We remember. Everywhere. This plugin gives your OpenClaw agents persistent, graph-connected memory across WhatsApp, Telegram, Discord, Slack, and every channel OpenClaw supports. All data stays on your machine.

Requirements

• Nowledge Mem desktop app or nmem CLI
• OpenClaw >= 2026.1.29

Installation

openclaw plugins install @nowledge/openclaw-nowledge-mem

Local mode (default)

Start Nowledge Mem desktop app or run nmem serve, then configure:

{
  "plugins": {
    "slots": { "memory": "openclaw-nowledge-mem" },
    "entries": {
      "openclaw-nowledge-mem": {
        "enabled": true
      }
    }
  }
}

That's it. The agent gets 7 tools and calls them on demand. No extra tokens wasted.

Remote mode

Connect to a Nowledge Mem server running elsewhere — on a VPS, a home server, or shared team instance. See remote access guide for server setup.

{
  "plugins": {
    "slots": { "memory": "openclaw-nowledge-mem" },
    "entries": {
      "openclaw-nowledge-mem": {
        "enabled": true,
        "config": {
          "apiUrl": "https://nowledge.example.com",
          "apiKey": "your-api-key-here"
        }
      }
    }
  }
}

The apiKey is injected as NMEM_API_KEY into the nmem CLI process — never passed as a CLI argument, never logged.

Tools

OpenClaw Memory Compatibility

These satisfy the OpenClaw memory slot contract and activate the "Memory Recall" section in OpenClaw's system prompt.

memory_search — Multi-signal recall using embedding, BM25, label match, graph signals, and recency decay. Returns structured source paths (nowledgemem://memory/<id>) for follow-up with memory_get or nowledge_mem_connections.

memory_get — Read a specific memory by ID or path. Supports MEMORY.md alias for Working Memory.

Nowledge Mem Native

These reflect capabilities unique to Nowledge Mem's knowledge graph architecture.

nowledge_mem_save — Capture structured knowledge with type classification, labels, and temporal context.

text:             "We decided to use PostgreSQL with JSONB for the task events table"
title:            "Task events database choice"
unit_type:        decision
importance:       0.8
labels:           ["backend", "architecture"]
event_start:      2024-03
temporal_context: past
→ Saved: Task events database choice [decision] (id: mem_abc) · labels: backend, architecture · event: 2024-03

Eight memory types: fact, preference, decision, plan, procedure, learning, context, event — each becomes a typed node in the knowledge graph. Labels enable filtering in memory_search. event_start records when something happened, not just when you saved it — powering bi-temporal search.

nowledge_mem_context — Read today's Working Memory: focus areas, priorities, unresolved flags, and recent activity. Generated by the Knowledge Agent each morning, updated throughout the day.

nowledge_mem_connections — Explore the knowledge graph around a topic or memory. Returns connected memories, EVOLVES version chains (how understanding has grown), related entities, and source document provenance (which files or URLs knowledge was extracted from).

memoryId: "mem_abc"
→ Connected memories:
  - PostgreSQL optimization patterns: Use JSONB GIN indexes for...
  - Redis caching layer decision: For frequently accessed task lists...
  Source documents (provenance):
  - api-spec.pdf (file): API specification for task management...
  Related entities:
  - PostgreSQL (Technology)
  - Task Management API (Project)
  Knowledge evolution:
  - superseded by newer understanding (version chain)

nowledge_mem_timeline — Browse your knowledge history chronologically. Use for questions like "what was I working on last week?" or "what happened yesterday?". Groups activity by day: memories saved, documents ingested, insights generated, and more.

last_n_days: 7
→ 2026-02-18:
  - [Memory saved] UV guide — Python toolchain setup
  - [Knowledge extracted from document] api-spec.pdf
2026-02-17:
  - [Daily briefing] Focus: NebulaGraph, AI biotech...
  - [Insight] Connection between Redis caching and...

nowledge_mem_forget — Delete a memory by ID or search query. Supports user confirmation when multiple matches are found.

Operating Modes

The plugin supports three modes. The default (tool-only) gives the agent full access to all tools with zero token overhead.

| Mode | Config | Behavior | |------|--------|----------| | Tool-only (default) | autoRecall: false, autoCapture: false | Agent calls tools on demand. Zero overhead. | | Auto-recall | autoRecall: true | Working Memory + relevant memories injected at session start. | | Auto-capture | autoCapture: true | Thread capture + LLM distillation at session end. |

Auto-Recall (autoRecall, default: false)

When enabled, the plugin injects Working Memory and relevant search results at session start. Useful for giving the agent immediate context without waiting for it to search proactively.

Auto-Capture (autoCapture, default: false)

When enabled, two things happen at agent_end, after_compaction, and before_reset:

1. Thread capture (always). The full conversation is appended to a persistent thread. Unconditional, idempotent by message ID.

2. LLM distillation (when worthwhile). A lightweight LLM triage determines if the conversation has save-worthy content. If yes, a full distillation pass creates structured memories with types, labels, and temporal data. Language-agnostic — works in any language.

Slash Commands

| Command | Description | |---------|-------------| | /remember <text> | Save a quick memory | | /recall <query> | Search your knowledge base | | /forget <id or query> | Delete a memory |

CLI Commands

openclaw nowledge-mem search "database optimization"
openclaw nowledge-mem status

Configuration

| Key | Type | Default | Description | |-----|------|---------|-------------| | autoRecall | boolean | false | Inject Working Memory + relevant memories at session start | | autoCapture | boolean | false | Thread capture + LLM distillation at session end | | maxRecallResults | integer | 5 | Max memories to recall at session start (only used when autoRecall is enabled) |

What Makes This Different

• Local-first: no API key, no cloud account. Your knowledge stays on your machine.
• Knowledge graph: memories are connected nodes, not isolated vectors. EVOLVES edges track how understanding grows over time.
• Source provenance: the Library ingests PDFs, DOCX, URLs — extracted knowledge links back to the exact document section it came from.
• Working Memory: an AI-generated daily briefing that evolves — not a static user profile.
• Cross-AI continuity: knowledge captured in any tool (Cursor, Claude, ChatGPT) flows to OpenClaw and back.
• Typed memories: 8 knowledge types mapped to graph node properties — structured understanding, not text blobs.
• Multi-signal search: not just semantic similarity — combines embedding, BM25 keyword, label match, graph & community signals, and recency/importance decay. See Search & Relevance.

License

MIT